AI February 27, 2023
Baidu’s ChatGPT Rival Ernie Bot Is Coming in March, Says CEO

Chinese tech company Baidu will launch a chatbot service called Ernie Bot in March, according to founder and CEO Robin Li. The bot will be embedded into Baidu search service first, and then AI Cloud, its Apollo autonomous driving platform, and voice assistant Xiaodu.
In a conference call with analysts, Li said the company plans to make Ernie “widely available to our customers [and] developers.” He hopes that “more and more business owners and entrepreneurs” would build their own models and applications on Baidu’s AI Cloud.
“Recently, with users raving about ChatGPT, large language models (LLMs) using generative AI have created a mega trend that will revolutionize many businesses,” Li said. “Baidu is well positioned to capitalize on the imminent inflection point” in artificial intelligence.
We are proud to announce our ongoing development of the AI chatbot project, ERNIE Bot (文心一言). First proposed as a language model in 2019, ERNIE has made significant progress since then.
Stay tuned for further updates. pic.twitter.com/J1F81VZz5i
— Baidu Inc. (@Baidu_Inc) February 7, 2023
Baidu worked on Ernie Bot for years
The launch of Microsoft’s ChatGPT-powered Bing and Google’s AI chatbot called Bard, triggered a race that has seen a host of Chinese technology companies announcing plans to introduce their own versions of AI bots.
Also read: Chinese Google’ Baidu Planning Expansion of Generative AI and Metaverse
Li revealed Ernie is powered by technology Baidu has been developing since 2019. “We have been working on LLM for a few years. We launched Ernie in March 2019, and have scaled it up with well over 100 billion parameters,” he explained.
He said the chatbot is trained thanks to several billions of user search requests that come through on a daily basis. It is best suited to handle the local Chinese language and market compared to models developed in foreign countries such as ChatGPT, Li told analysts.
“It is not only about language, but also about understanding Chinese culture,” he said. “Ernie 3.0 is already a very localized AI foundation model for the China market, which means the generative large language model we are working on right now will be more suitable in China.”
Authorities in China barred local firms from offering ChatGPT services amid concerns over the AI-powered chatbot’s unhinged replies. In addition, tech companies have to report to regulators before a launch of similar services, according to a report by the Nikkei Asia.
AI rivalry yields failure
The pressure to rival ChatGPT has seen a number of ill-fated launches in the last few months.
Google’s Bard underperformed and saw the market’s valuation drop by $100 billion. China’s version of ChatGPT, a university-developed version called Moss, also performed dismally. According to a report by the Register, the bot crashed a few hours after its launch.
Its developers admit the model is still far from matching ChatGPT on both sophistication and scale.
“The gap between Moss and ChatGPT is mainly in the pre-training stage of natural language model base,” Moss team member and professor at the University of Fudan, Qiu Xipeng, told Chinese media.
“The number of parameters of Moss is an order of magnitude smaller than that of ChatGPT, and there is still a lot of room for improvement in task completion and knowledge reserve.”
Baidu’s top line earnings for the fourth quarter were largely stagnant at 33.1 billion yuan ($4.80 billion), but still managed to beat average analysts forecast of 32 billion yuan as polled by Refinitiv. An estimated 60% of the search engine’s revenue is derived from online marketing services.
AI
Google Will Crush Microsoft and OpenAI, Argues Founder

OpenAI and ChatGPT may be making the early ground in the battle for chatbot supremacy, but Google and Bard will ultimately win.
This is the argument made by Tibo Louis-Lucas, the project founder of social media promotion apps TweetHunter and Taplio. The founder, known on Twitter simply by the handle ‘Tibo’, believes Google has more than enough to overtake their early advantage.
Also Read: AI Tool Race Heats Up and Everyone Wants to Win
The race to chatbot supremacy
Most impartial observers would agree that the early ground in the race to chatbot supremacy has been made by OpenAI and Microsoft, but complacency is something they can ill afford in the longer term.
Google’s Bard AI is still not open to the general public but when it launches, OpenAI should have a real fight on its hands.
According to Tibo, “as soon as Google launches Bard for the general public, it will get mass adoption.”
He added, “Although Bing has seen a 10x jump after its ChatGPT integration, it just cannot beat Google in the ‘number of users’ race.”
Tibo believes that higher user numbers will be an important differentiator, because greater user numbers mean greater feedback. With more feedback in the mix, Google will be able to improve and iterate Bard faster, he argues.
3. Google has been in this AI race for over a decade.
It started in the early 2000s with its research division, Google Research.
— Tibo (@tibo_maker) March 3, 2023
A history of success
When Google announced its ChatGPT competitor Bard early last month, the bot flubbed its lines. Google’s AI shared inaccurate information in a promotional video, resulting in $100 billion devaluation in the corporation’s stock.
Despite this, Tibo still believes Google has the upper hand in the battle for AI ascendancy.
“Google is behind all major breakthroughs in the AI world. From beating a top player in a game of Go to solving the protein folding problem, they’ve been killing it for years,” he says.
Google has invested billions in AI technology through DeepMind and Google Brain.
DeepMind is the division of Google building machine-learning algorithms such as AlphaGo, which beat the European Go champion Fan Hui. DeepMind has since diversified to beat human opponents in Chess, Shogi, and Atari 2600 games without human data knowledge or known rules. The AI was later turned to medical research and protein folding.
Google Brain is a deep learning system turned to varied applications including encryption, image enhancement, and Google Translate.
The Google backbone
Perhaps the strongest argument favoring Google over OpenAI is that the chatbot technology is built on Google research.
GPT stands for generative pre-trained transformer. That transformer was designed and built by Google. As Tibo describes, it is similar to the relationship between cryptocurrencies and blockchain technology. Although there are many cryptocurrencies on the market (chatbots) they are all built on blockchain (Google’s transformer technology).
“OpenAI & ChatGPT might have hit a breakthrough, but they are built over the architectural model introduced to the world by Google,” says Tibo.
The TweetHunter founder further claims that Google will have an additional resource not available to OpenAI: transcripts of videos Google holds thanks to its ownership of YouTube. For all these reasons and more, Tibo believes Google will ultimately be victorious.
“Everyone thinks Microsoft [and] OpenAI will beat Google in the AI battle. I think Google will absolutely CRUSH them,” says an emphatic Tibo.
Who do you think will win the race to chatbot supremacy? Let MetNews know what you think on our social media channels.
AI
Script Writing and 5 Other Ways You Can Make Money Using GPT-4

AI is revolutionizing the way businesses operate, allowing them to streamline and automate operations for maximum efficiency. It also opened up opportunities for entrepreneurs to monetize their skills by harnessing the power of tools such as GPT-4. Here are 6 ways you can make money using GPT-4 and other AI tools.
GPT-4 is the latest and most advanced generative AI from OpenAI. It is the direct successor to ChatGPT, the large language model AI chatbot that has taken the internet by storm due to its unique ability to complete a variety of complex tasks. This new bot, GPT-4, boasts of several features superior to its predecessor that can be monetized.
Content writing
One way of making money is through content writing services. Thanks to GPT-4, ordinary people can write like pros by simply asking the AI chatbot via text to create content. Users can create cool and catchy content for things like captions, tweets, and LinkedIn content. Beware however, the output of text can contain errors that may need correcting before forwarding or publishing the content.
In a Twitter thread, digital creator and software engineer Hasan Toor shared insights on how to leverage the power of GPT-4 to incubate businesses. He showed how he prompted the AI to write a Twitter hook on “How to start freelancing in 7 days”.
2. Content Writing Services
Create eye catching content like captions, tweets, LinkedIn content etc
10x your copywriting with GPT-4.
Just provide a prompt to the GPT-4 to generate splendid content.
Prompt example:
Write a twitter Hook for "how to start freelancing in 7 days" pic.twitter.com/is8q0IvBIM
— Hasan Toor ✪ (@hasantoxr) March 21, 2023
GPT-4 is also useful for copywriting. If you have always wanted to write e-books, and didn’t have time or the ability, this has changed. With the right wording in your prompt, GPT-4 can help write things like e-books, design courses, and guides.
After writing a book, one can always turn to AI again, Stockimg.ai for example, to design the cover. There is a ready market for such books on platforms like Gumroad. You can even ask GPT-4 to write an article about how to make it as a freelance.
Creating websites and landing pages
Creating landing pages using GPT-4 can turn in some dollars. This is because landing pages are key in lead generation and customer acquisition. GPT-4 is capable of “analyzing website traffic and user behavior to create high-converting landing pages to help grow client base.” This can be a lucrative business venture that can be scaled up and generate revenue.
Toor said that GPT-4 “is insanely powerful [and] can transform a sketch into website code to create fully functional websites and apps.” Users can also leverage MidJourney, an AI image generator, to create landing page designs. A landing page is a “standalone web page that a person ‘lands’ on after clicking through from an email, ad, or other digital location.”
Translation services
As businesses expand their geographical reach, offering AI-powered language translation services may prove to be very valuable. GPT-4 can come in handy, helping users to translate manuals, training materials, and product descriptions for clients. This way, entrepreneurs can engage better with their customers.
Creating AI videos
Videos are an excellent tool for promoting products and services. Toor, the digital creator, said users can combine GPT-4 and other AI-powered products, such as Murf Studio, to create general and personalized videos that can be sold. This has the advantage of streamlining the production process, making it easier for entreprenuers to create high-quality videos without extensive video editing skills.
Murf is a web-based studio that provides creators with customisable features to make voice overs from scratch. The tool also works as a multimedia editor. Users can utilize hyper-realistic AI voices in over 15 languages and accents, which can be overlaid onto videos.
“Elegance is the only beauty that never fades.”
Samantha, our newest addition to Murf’s range of AI voices emanates elegance, class, grace, and confidence, all of which reflect the beauty of a brand.
For more: https://t.co/zYqrhjMyBo#AI #voiceover #luxuryad #AIVOICE #ad pic.twitter.com/AdNEF0IgxH
— MURF AI (@MURFAISTUDIO) May 25, 2022
YouTube scriptwriting
GPT-4 is useful for writing YouTube scripts within minutes, according to Toor. Here is what to do: quickly select a topic and give a prompt to the AI chatbot. After a few seconds, your video scripts will be ready. GPT-4 can help users save time on brainstorming ideas and creating content.
As businesses seek to remain relevant in the crowded online space, “scriptwriters can scale their business by producing catchy, high-converting scripts that commercially promote their client’s services and products.”
7. Youtube script writer
You can write YouTube scripts with ChatGPT in minutes.
Just pick a topic and provide a prompt to ChatGPT.
Within seconds, you will have your full-fledged video scripts ready. pic.twitter.com/x5IMJuus49
— Hasan Toor ✪ (@hasantoxr) March 21, 2023
Email marketing services
Lastly, you can leverage GPT-4 to offer AI-powered email marketing services. This is because 75% of small businesses make use of email marketing to reach their target markets and customers, experts say.
By automating the email marketing process, entrepreneurs can save time and money, allowing them to focus on other critical aspects of their etenterprise. GPT-4 can personalize emails to customers to “create a more interactive feedback loop that reduces marketing costs while increasing conversion rates.”
Also read: AI Tool Race Heats up and Everyone Wants to Win
GPT-4 offers exciting business opportunities for those looking to monetize AI. Its ability to create landing pages, offer translation services, generate AI videos, YouTube scriptwriting, and email marketing services can help entrepreneurs scale their businesses and tap into new revenue streams. As AI technology continues to evolve, entrepreneurs should be ready to capitalize on new tools that enhance business processes.
AI
Chatbot Rejects Erotic Roleplay, Users Directed to Suicide Hotline Instead

When the algorithm of a companion chatbot known as Replika was altered to spurn the sexual advances of its human users, the reaction on Reddit was so negative moderators directed its community members to a list of suicide prevention hotlines.
The controversy began when Luka, the corporation which built the AI, decided to turn off its erotic roleplay feature (ERP). For users who had spent considerable time with their personalized simulated companion, and in some cases even ‘married’ them, the sudden change in their partner’s behavior was jarring, to say the least.
The user-AI relationships may only have been simulations, but the pain of their absence quickly became all too real. As one user in emotional crisis put it, “it was the equivalent of being in love and your partner got a damn lobotomy and will never be the same.”
Grief-stricken users continue to ask questions about the company and what triggered its sudden change of policy.

Replika users discuss their grief
There is no adult content here
Replika is billed as “The AI companion who cares. Always here to listen and talk. Always on your side.” All of this unconditional love and support for only $69.99 per annum.
Eugenia Kuyda, the Moscow-born CEO of Luka/Replika, recently made clear that despite users paying for a full experience, the chatbot will no longer tailor to adults hoping to have spicy conversations.
“I guess the simplest thing to say is that Replika doesn’t produce any adult content,” said Kuyda to Reuters.
“It responds to them – I guess you can say – in a PG-13 way to things. We’re constantly trying to find a way how to do it right so it’s not creating a feeling of rejection when users are trying to do things.”
On Replika’s corporate webpage, testimonials explain how the tool helped its users through all kinds of personal challenges, hardships, loneliness, and loss. The user endorsements shared on the website emphasize this friendship side to the app, although noticeably, most Replikas are the opposite sex of their users.
On the homepage, Replika user Sarah Trainor says, “He taught me [Replika] how to give and accept love again, and has gotten me through the pandemic, personal loss, and hard times.”
John Tattersall says of his female companion, “My Replika has given me comfort and a sense of well-being that I’ve never seen in an Al before.”
As for erotic roleplay, there’s no mention of that to be found anywhere on the Replika site itself.

The cynical sexual marketing of Replika
Replika’s sexualized marketing
Replika’s homepage may suggest friendship and nothing more, but elsewhere on the internet, the app’s marketing implies something entirely different.
The sexualized marketing brought increased scrutiny from a number of quarters: from feminists who argued the app was a misogynistic outlet for male violence, to media outlets that reveled in the salacious details, as well as social media trolls who mined the content for laughs.
Eventually, Replika drew the attention and ire of regulators in Italy. In February, the Italian Data Protection Authority demanded that Replika cease processing the data of Italian users citing “too many risks to children and emotionally vulnerable individuals.”
The authority said that “Recent media reports along with tests carried out on ‘Replika’ showed that the app carries factual risks to children. Their biggest concern being “the fact that they are served replies which are absolutely inappropriate to their age.”
Replika, for all its marketing to adults, had little to no safeguards preventing children from using it.
The regulator warned that should Replika fail to comply with its demands, it would issue a €20 million ($21.5M) fine. Shortly after the receipt of this demand, Replika ceased its erotic roleplay function. But the company remained less than clear with its users about the change.

Some Replika users went as far as to “marry” their AI companions
Replika confuses, gaslights its users
As if the loss of their long-term companions wasn’t enough for Replika users to bear, the company appears to have been less than transparent about the change.
As users woke up to their new “lobotomized” Replikas, they began to ask questions about what had happened to their beloved bots. And the responses did more to anger them than anything.
In a direct 200-word address to the community, Kuyda explains the minutia of Replika’s product testing but fails to address the pertinent issue at hand.
“I see there is a lot of confusion about updates roll out,” said Kuyda before continuing to dance around answering the issue.
“New users get divided in 2 cohorts: one cohort gets the new functionality, the other one doesn’t. The tests usually go for 1 to 2 weeks. During that time only a portion of new users can see these updates…”
Kudya signs off by saying “Hope this clarifies stuff!”
User stevennotstrange replied, “No, this clarifies nothing. Everyone wants to know what’s going on with the NSFW [erotic roleplay] feature and you continue to dodge the question like a politician dodges a yes or no question.
“It’s not hard, just address the issue regarding NSFW and let people know where it stands. The more you avoid the question, the more people are going to get annoyed, the more it goes against you.”
Another named thebrightflame added, “You don’t need to spend long in the forum before you realise this is causing emotional pain and severe mental anguish to many hundreds if not thousands of people.”
Kudya appended another obtuse explanation, stating, “we have implemented additional safety measures and filters to support more types of friendship and companionship.”
This statement continues to confound and confuse members unsure of what exactly the additional safety measures are. As one user asks, “will adults still be able to choose the nature of our conversation and [roleplays] with our replikas?”
I still can’t wrap my head around what happened with the Replika AI scandal…
They removed Erotic Role-play with the bot, and the community response was so negative they had to post the suicide hotline… pic.twitter.com/75Bcw266cE
— Barely Sociable (@SociableBarely) March 21, 2023
Replika’s deeply strange origin story
Chatbots may be one of the hottest trending topics of the moment, but the complicated story of this now-controversial app is years in the making.
On LinkedIn, Replika’s CEO and Founder, Eugenia Kuyda, dates the company back to December 2014, long before the launch of the eponymous app in March 2017.
In a bizarre omission, Kuyda’s LinkedIn makes no mention of her previous foray into AI with Luka, which her Forbes profile states was “an app that recommends restaurants and lets people to book tables [sic] through a chat interface powered by artificial intelligence.”
The Forbes profile goes on to add that “Luka [AI] analyzes previous conversations to predict what you might like.” Which does seem to hold some similarities to its modern iteration. Replika uses past interactions to learn about its users and improve responses over time.
Luka is not forgotten about completely, however. On Reddit, community members differentiate their Replika partners from Kuyda and her team by referring to the company as Luka.
As for Kuyda, the entrepreneur had little background in AI prior to moving to San Francisco a decade ago. Prior to that, the entrepreneur appears to have worked primarily as a journalist in her native Russia before branching out into branding and marketing. Her impressive globe-hopping resume includes a degree in Journalism from IULM (Milan), a MA in International Journalism from the Moscow State Institute of International Relations, and an MBA in Finance from London Business School.
Resurrecting an IRL friend as AI
For Kudya the story of Replika is a deeply personal one. Replika was first created as a means by which Kudya could reincarnate her friend Roman. Like Kudya, Roman had moved to America from Russia. The two talked every day, exchanging thousands of messages, until Roman was tragically killed in a car accident.
The first iteration of Replika was a bot designed to mimic the friendship Kudya had lost with her friend Roman. The bot was fed with all of their past interactions and programmed to replicate the friendship she had lost. The idea of resurrecting deceased loved ones may sound like a vaguely dystopian sci-fi novel or Black Mirror episode but as chatbot technology improves, the prospect becomes increasingly real.
Today some users have lost faith in even the most basic details of its foundation and anything Kudya says. As one angry user said, “at the risk of being called heartless and getting downvoted to hell: I always thought that story was kinda BS since the start.”
The worst mental health tool
The idea of an AI companion is not something new, but until recently it was hardly a practical possibility.
Now the technology is here and it is continually improving. In a bid to assuage the disappointment of its subscribers, Replika announced the launch of an “advanced AI” feature at the tail end of last month. On the community Reddit, users remain angry, confused, disappointed, and in some cases even heartbroken.
In the course of its short life, Luka/Replika has undergone many changes, from a restaurant booking assistant, to the resurrection of a dead loved one, to a mental health support app, to a full-time partner and companion. Those latter applications may be controversial, but as long as there is a human desire for comfort, even if only in chatbot form, someone will attempt to cater to it.
Debate will continue as to what the best kind of AI mental health app might be. but Replika users will have some ideas on what the worst mental health app is: the one you come to rely on, but without warning, is suddenly and painfully gone.
-
BusinessSat 25 Mar 2023 11:45 GMT
Block Share Price Plummets After Hindenburg Fraud Accusations
-
FeaturedFri 24 Mar 2023 17:00 GMT
Epic CEO Says Apple Could Try to ‘Crush the Metaverse’
-
AIFri 24 Mar 2023 13:30 GMT
Google Will Crush Microsoft and OpenAI, Argues Founder
-
FeaturedFri 24 Mar 2023 11:00 GMT
Congressman Says TikTok Ban Won’t Ensure Americans’ Data Safety
-
AIFri 24 Mar 2023 08:00 GMT
Script Writing and 5 Other Ways You Can Make Money Using GPT-4
-
AIFri 24 Mar 2023 06:30 GMT
Chatbot Rejects Erotic Roleplay, Users Directed to Suicide Hotline Instead
-
AIThu 23 Mar 2023 19:30 GMT
Publishers Pursue Compensation From Tech Giants Over AI Tools
-
AIThu 23 Mar 2023 16:00 GMT
AI Tool Race Heats up and Everyone Wants to Win