Posts

ai

President Biden Issues Executive Order for AI Oversight

On Monday, President Joe Biden signed an executive order covering a wide range of topics related to artificial intelligence, paving the way for new government regulations and funding.

The 111-page order covers multiple facets of AI and areas of concern or development, including civil rights, cybersecurity, discrimination, global competition, and a push for establishing federal AI jobs.

A senior White House official, who wished to remain anonymous, reportedly told NBC News that the potential uses of AI are so vast that effective regulations must cover a lot of ground. He also underscored the need for “significant bipartisan legislation.”

“AI policy is like running into a decathlon, and there’s 10 different events here, and we don’t have the luxury of just picking ‘we’re just going to do safety’ or ‘we’re just going to do equity’ or ‘we’re just going to do privacy.’ You have to do all of these things.”

Embed from Getty Images

The order expands on a July nonbinding agreement between seven of the most prominent U.S. technology companies developing AI. The agreement required the companies to hire outside experts to identify weaknesses in their systems. The government can legally require companies to disclose the results of those safety tests under the Defense Production Act.

The Department of Commerce will also be required to develop guidelines for properly “watermarking” AI content, such as “deepfake” videos and ChatGPT-generated essays.

In an interview with NBC News, the Stanford Institute for Human-Centered Artificial Intelligence co-director Fei-Fei Li stressed the importance of government funding for AI to solve society’s pressing issues.

“The public sector holds a unique opportunity in terms of data and interdisciplinary talent to cure cancer, cure rare diseases, to map out biodiversity at a global scale, to understand and predict wildfires, to find climate solutions, to supercharge our teachers. There’s so much the public sector can do, but all of this is right now starved because we are severely lacking in resources.”

Embed from Getty Images

Some of the other topics covered in the order are geared toward anticipating and mitigating real-world problems that may arise from the widespread implementation of AI.

For instance, it asks the Department of Labor to address the potential for AI to cause job losses; the Consumer Financial Protection Bureau and Department of Housing and Urban Development to address how AI may exacerbate discrimination in banking and housing sectors; and the Office of Management and Budget, and others, to determine how the government can use AI without jeopardizing privacy rights.

The AI Now Institute managing director, Sarah Myers West, praised President Biden for including ethical concerns in the executive order. The nonprofit focuses on the societal implications of artificial intelligence use.

“It’s great to see the White House set the tone on the issues that matter most to the public: labor, civil rights, protecting privacy, promoting competition. This underscores you can’t deal with the future risks of AI without adequately dealing with the present. The key to looking forward will be to ensure strong enforcement as companies attempt to set a self-regulatory tone: industry cannot be left to lead the conversation on how to adequately address the effects of AI on the broader public.”

ai

Amazon Invests up to $4 Billion in OpenAI Rival Anthropic in Exchange for Minority Stake

On Monday, Amazon announced it will invest up to $4 billion into the artificial intelligence company Anthropic. In exchange, Amazon will gain partial ownership, and Anthropic will use the company’s cloud computing platform, Amazon Web Services (AWS), more widely.

The growing relationship between the two firms is an example of how some large tech companies with extensive cloud computing resources are using those assets to strengthen their position in the artificial intelligence industry.

According to a statement released by Amazon, Anthropic will use AWS as its primary cloud provider, using the cloud platform to do most of its AI model development and research into AI safety. Anthropic will also have access to Amazon’s suite of in-house AI chips.

“AWS will become Anthropic’s primary cloud provider for mission-critical workloads, including safety research and future foundation model development. Anthropic plans to run the majority of its workloads on AWS, further providing Anthropic with the advanced technology of the world’s leading cloud provider.”

In addition, Anthropic has committed to making its AI models available to AWS users long-term, providing them with early access to features, including the ability to customize Anthropic models for their own purposes.

Embed from Getty Images

“With today’s announcement, customers will have early access to features for customizing Anthropic models, using their own proprietary data to create their own private models, and will be able to utilize fine-tuning capabilities via a self-service feature.”

Amazon Web Services (AWS) customers already have access to Anthropic’s AI models through Amazon Bedrock, the tech giant’s storefront for AI goods. Bedrock not only supports Amazon’s own models but also those from third-party developers such as  Stability AI and AI21 Labs.

In a press release, the co-founder and CEO of Anthropic, Dario Amodei, said that his company is “excited to use AWS’s Trainium chips to develop future foundation models.”

“Since announcing our support of Amazon Bedrock in April, Claude has seen significant organic adoption from AWS customers. By significantly expanding our partnership, we can unlock new possibilities for organizations of all sizes as they deploy Anthropic’s safe, state-of-the-art AI systems together with AWS’s leading cloud technology.”

Embed from Getty Images

Anthropic stated that Amazon’s minority stake would not alter the company’s corporate governance structure or its dedication to the ethical advancement of artificial intelligence.

“Our corporate governance structure remains unchanged, with the Long Term Benefit Trust continuing to guide Anthropic in accordance with our Responsible Scaling Policy. As outlined in this policy, we will conduct pre-deployment tests of new models to help us manage the risks of increasingly capable AI systems.”

Several cloud market leaders, like Microsoft and now Amazon, have made investments into artificial intelligence technology. OpenAI, the company that developed ChatGPT, received $1 billion from Microsoft in 2019. Microsoft recently also invested $10 billion in OpenAI and is striving to integrate OpenAI’s technology into consumer-facing Microsoft products such as Bing.

This deal is Amazon’s most recent push into the artificial intelligence space to compete with industry leaders like Microsoft and Alphabet’s Google.

Geoffrey Hinton, ‘The Godfather Of A.I.’, Leaves Google And Warns Of Future Dangers Of A.I.

In 2012, Geoffrey Hinton and two of his graduate students from the University of Toronto created technology that has become the foundation of Artificial Intelligence systems used by some of the biggest tech companies in the world. Now, Hinton has left his job at Google and is warning many about the risks of AI technology, stating that he now regrets his life’s work.

samsung

Alphabet Stock Drops After Samsung Considers Switching Default Search to Bing

Samsung is considering ending its long-standing partnership with Google in favor of Bing as the default search engine for its devices. Alphabet, Google’s parent company, saw its stock drop by more than 3% in early trading Monday as a result of the news.

According to The New York Times, Google’s internal response to the possible change was “panic.” It is unclear whether Microsoft’s decision to incorporate ChatGPT into Bing was a driving factor in Samsung’s deliberation.

Since Google makes the bulk of its money from search ads, the development of AI search technology represents the greatest threat to Google’s search business in the past 25 years. As it stands, Google’s contract with Samsung generates around $3 billion in annual revenue. Google’s contract with Apple, which is up for renewal this year, brings in $20 billion in revenue.

Recent advances in AI have prompted Google to explore ways to incorporate AI into its flagship search product. According to internal reports, around 160 designers, engineers, and executives at Google are all working together in “sprint rooms” on a new project called Magi to build competitive AI-powered search features.

Embed from Getty Images

Google is also working on a brand-new search engine that uses artificial intelligence more broadly. Plans for this new search engine are still in the early stages, and a release date has not been set. The future system will anticipate user searches, provide curated lists of potential purchases, and be more conversational, similar to Bing’s new search engine format. In the meantime, the goal of the Magi project is to enhance Google’s current search engine.

Jim Lecinski, a former Google vice president of sales and service and professor of marketing at Northwestern University, said the company has to now persuade users that it is as “powerful, competent and contemporary” as its competitors.

“If we are the leading search engine and this is a new attribute, a new feature, a new characteristic of search engines, we want to make sure that we’re in this race as well.”

During a test last week, Google employees quizzed the Magi project’s enhanced Google search on its conversational skills by asking it follow-up questions. Next month, Google plans to roll out the tools to the public, making them available to as many as 1 million users, with added functionality coming in the fall. According to The New York Times, the number will increase to 30 million users by the end of the year. The tools will only be available in the United States.

In a roadmap document, a Google executive revealed that the company is thinking about integrating artificial intelligence into Google Earth’s mapping features and providing a separate feature that lets users conduct music searches via conversation with a chatbot.

Embed from Getty Images

A feature that uses AI to generate images within image results is also in the works, as is a feature that uses AI text conversations to teach users a new language. Users will also be able to ask a chatbot questions as they browse the web using a product called Searchalong.

Last year, Google made $162 billion from its search business. According to The New York Times, Google was taken aback by the prospect of Samsung, which sells hundreds of millions of smartphones with Google’s Android software every year, switching its default search engine.

In a statement to CNN, Google spokesperson Lara Levin said the company has always been committed to using AI to “improve the quality of our results” and “offer entirely new ways to search,” citing a feature rolled out last year that allowed users to search by combining images and words.

“We’ve done so in a responsible and helpful way that maintains the high bar we set for delivering quality information. Not every brainstorm deck or product idea leads to a launch, but as we’ve said before, we’re excited about bringing new AI-powered features to search and will share more details soon.”

Google has been involved in AI research for a long time, and its DeepMind lab in London is world-renowned. The company has also contributed to the development of autonomous vehicles and large language models used by chatbots. Due to concerns over the accuracy of AI, Google has been slow to integrate it into its search engine.

snapchat

Snapchat Launches AI Chatbot Powered by ChatGPT

Snapchat is launching its own artificial intelligence chatbot powered by OpenAI’s viral ChatGPT. The feature, called My AI, will be available to Snapchat Plus subscribers starting this week.

In a blog post, Snapchat shared how My AI can help subscribers with various tasks and assist them in their day-to-day activities.

“My AI can recommend birthday gift ideas for your BFF, plan a hiking trip for a long weekend, suggest a recipe for dinner, or even write a haiku about cheese for your cheddar-obsessed pal.”

Snapchat cautioned, however, that the chatbot is “experimental” and may respond in unexpected ways.

“As with all AI-powered chatbots, My AI is prone to hallucination and can be tricked into saying just about anything. Please be aware of its many deficiencies, and sorry in advance!”

An AI hallucination is a term for when an AI presents false facts as the truth. In other words, it may confidently output completely made-up answers, leading to misinformation. At times, the answers may even be nonsensical.

Embed from Getty Images

In an email to CNET, a Snapchat representative described how the company customized the latest version of OpenAI’s ChatGPT technology for its platform.

“My AI was trained to have a unique tone and personality that plays into Snapchat’s core values around friendship, learning, and fun. It has been trained to adhere to our trust and safety guidelines.”

The company’s community guidelines prohibit the chatbot from responding with explicit, inflammatory or violent content.

Snapchat will store all conversations between subscribers and their My AI to review and help improve product experience. Users can also submit direct feedback to Snapchat by pressing and holding any message. Snapchat advised users not to “share any secrets with My AI” and to not rely on it for advice.

Currently, the feature is only available to Snapchat Plus members. However, in an interview with the Verge, Snapchat founder and CEO Evan Spiegel said the goal is to make the feature available to all of Snapchat’s 750 million monthly users.

Embed from Getty Images

Snapchat is the latest in a string of companies to integrate artificial intelligence into their platforms. Google recently revealed its ChatGPT contender Bard. The chatbot infamously made a factual error during an ad demo, costing Google a $100 billion drop in market value.

The same week, Microsoft announced it would integrate ChatGPT into its search engine Bing.

Since Snapchat is a messaging service, Spiegel believes it is uniquely positioned to create a personable chatbot. Spiegel told the Verge, “The big idea is that in addition to talking to our friends and family every day, we’re going to talk to AI every day.”

Unlike other AI chatbot integrations on platforms, Snapchat’s My AI interface suggests the chatbot is intended to be more than just a productivity tool. The chatbot has an avatar, and its “user profile” resembles a regular Snapchat friend profile. Users can even change the chat’s wallpaper.

The current price for a subscription to Snapchat Plus is $3.99 per month.

bard

Google’s AI Chatbot ‘Bard’ Makes a Factual Error, Costing Alphabet $100 Billion in Market Value

Alphabet, the parent company of Google, saw a $100 billion drop in market value after Bard, its competing chatbot to Microsoft’s ChatGPT, made an error in an ad containing a demo.

ai

Real Estate Agents Utilizing ChatGPT AI Tool To Generate Listings

Real estate agents across the nation are utilizing ChatGPT, an AI chatbot tool that can generate detailed descriptions based on prompts given by the user.