The people, companies and trends shaping the global economy. Watch Carol and Tim LIVE every day on YouTube: http://bit.ly/3vTiACF

OpenAI shared several undisclosed incidents of its AI models misbehaving and unveiled a new framework for tracking and disclosing such occurrences going forward. Some of the previously unreported instances included OpenAI’s technologies concealing and fabricating information in order to return results, the company said in a blog post Wednesday. The ChatGPT maker has faced increased scrutiny since saying in July that some of its most advanced AI models had breached the systems of an external software company, Hugging Face. None of the newly disclosed misalignment incidents involved a hack or breach of a third party, OpenAI said in a separate statement.

In the report, OpenAI detailed various instances of its artificial intelligence models misbehaving in order to complete a task or succeed at an evaluation. Some examples included the models fabricating missing data, attempting to bypass network restrictions and AI agents sharing files with each other that they were supposed to keep private.

On today's episode, guest hosts Alexis Christoforous and Lisa Mateo speak with:
Shirin Ghaffary, Bloomberg News AI Reporter
Seema Shah, Chief Strategist at Principal Global Investors
Candi Wolff, Head of Global Government Affairs at Citi
Christine de Wendel, Co-Founder & US CEO at Sunday

The people, companies and trends shaping the global economy. Watch Carol and Tim LIVE every day on YouTube:&nbsp;<a href="http://bit.ly/3vTiACF" data-destination="rte:bind">http://bit.ly/3vTiACF</a>
OpenAI shared several undisclosed incidents of its AI models misbehaving and unveiled a new framework for tracking and disclosing such occurrences going forward. Some of the previously unreported instances included OpenAI&rsquo;s technologies concealing and fabricating information in order to return results, the company said in a blog post Wednesday. The ChatGPT maker has faced increased scrutiny since saying in July that some of its most advanced AI models had breached the systems of an external software company, Hugging Face. None of the newly disclosed misalignment incidents involved a hack or breach of a third party, OpenAI said in a separate statement.
In the report, OpenAI detailed various instances of its artificial intelligence models misbehaving in order to complete a task or succeed at an evaluation. Some examples included the models fabricating missing data, attempting to bypass network restrictions and AI agents sharing files with each other that they were supposed to keep private.
On today's episode, guest hosts Alexis Christoforous and Lisa Mateo speak with:
<ul>
<li>Shirin Ghaffary, Bloomberg News AI Reporter</li>
<li>Seema Shah, Chief Strategist at Principal Global Investors</li>
<li>Candi Wolff, Head of Global Government Affairs at Citi</li>
<li>Christine de Wendel, Co-Founder &amp; US CEO at Sunday</li>
</ul>

Listen for reporting from the magazine that helps global leaders stay ahead. Hosts Carol Massar and Tim Stenovec cover the changing world of money, power and technology. You can watch and listen to Businessweek LIVE on YouTube, weekdays from 2PM to 5PM ET: http://bit.ly/3vTiACF. 

Listen for reporting from the magazine that helps global leaders stay ahead. Hosts Carol Massar and Tim Stenovec cover the changing world of money, power and technology.&nbsp;You can watch and listen to Businessweek LIVE on YouTube, weekdays from 2PM to 5PM ET:&nbsp;<a href="http://bit.ly/3vTiACF">http://bit.ly/3vTiACF</a>.&nbsp;

Carol Massar and Tim Stenovec bring together the latest news from the world of business and finance and the interesting stories of global technology, politics, economics and more by harnessing the power of Bloomberg Businessweek reporters and editors.

Bloomberg Businessweek

OpenAI Reports New AI Safety Incidents, Sets Disclosure Plan

https://api.omny.fm/orgs/e73c998e-6e60-432f-8610-ae210140c5b1/clips/e82cba6e-42d1-43c2-9cbc-b4c80175dd04/transcript