• Bytes of AI
  • Posts
  • U.S Takes Steps Toward AI Regulation, Apple Mum on LLM, AI Golf, More

U.S Takes Steps Toward AI Regulation, Apple Mum on LLM, AI Golf, More

Top Byte: U.S. Takes Steps Toward AI Regulation

As expected, on Monday President Biden issued an executive order to create some AI safeguards. 

Provisions of the order include a requirement for AI companies to report if their models could be used by foreign governments or terrorists to help create weapons of mass destruction, attempts to identify deep fakes via watermarks, direction for the government to create standards to combat bias in AI, and a loosening visa restrictions to bring more AI-related tech talent into the country. Executives from companies like Microsoft, Alphabet and Amazon were present at the signing.

Just two later, the U.S. joined 27 other countries at Britain's AI Safety summit in pledging to work together to fight the "inherit risks" of AI. 

What doesn't appear to have been on the agenda at these events is just how feasible any regulation may be. For example, many studies -- including this one released last month -- have shown that watermarking AI-generated images isn't a solution that holds up in real-world environments. 

 And there's not just technical challenges: The Brookings Institute offers an excellent in-depth look at many of the issues around regulating AI here.  

More Generative AI News This Week: 

  • Earlier this week Apple CEO Tim Cook talked loosely about the company's ongoing AI efforts, but offered no comment on what some people really wanted to hear about: details on the company's rumored LLM projects (some reports say there are several in the works). 

  • In contrast, Elon Musk is talking about his company’s Chatbot, announcing Friday that XAI has released one in a restricted preview. 

  • OpenAI is rolling out new changes to ChatGPT Plus, including integrated PDF interaction (before, users needed a plug in) and more memory for longer input. Also, the training data cut-off for ChatGPT-4 now appears to be April 2023. 

  • Speaking of OpenAI, the company will be hosting its first ever live event (aimed at developers) on Monday. It's sold out, but interest is so high some people are forming virtual meetups or even live get togethers to watch the keynote (it will be available for streaming here).

  • The healthcare industry is looking into the ethics of patients interacting with AI technology. 

  • And yes, AI is indeed coming to everything -- this week it's golf putters.

RESOURCE OF THE WEEK:
AI in the Enterprise Summit

February 19-21 in Las Vegas, The TDWI Institute is holding a live conference for executives, TDWI Modern Data Leader's Summit: AI in the Enterprise.

The event’s Web site says it’s “an intensive, 2.5-day event for data and business leaders seeking to harness the transformative power of AI within their organizations.” The full agenda can be found here.

Attendees are eligible to earn a “TDWI Modern Data Leadership Certificate on AI in the Enterprise.”

For more information, go here.