[1/4] The Google, Microsoft and Alphabet logos and the words AI Artificial Intelligence can be seen in this image taken on May 4, 2023. REUTERS/Dado Ruvic/Illustration/File Photo
SAN FRANCISCO, June 15 (Reuters) – Alphabet Inc (GOOGL.O) is warning employees about how they use chatbots, including its own Bard, as it markets the program around the world, four people familiar with the issue Reuters.
Google’s parent company has advised employees not to enter its confidential material into AI chatbots, the people said, and the company confirmed this, citing a long-standing information protection policy.
The chatbots, including Bard and ChatGPT, are human-sounding programs that use so-called generative artificial intelligence to hold conversations with users and answer countless prompts. Human reviewers are allowed to read the chats, and researchers found that similar AI could reproduce the data it recorded during training, creating a leak risk.
Alphabet also warned its engineers to avoid direct use of computer code that chatbots can generate, some people said.
When asked for comment, the company said Bard can make unwanted code suggestions, but it nevertheless helps programmers. Google also said it wanted to be transparent about the limitations of its technology.
The concerns show how Google wants to prevent business harm from software it launched in competition with ChatGPT. In Google’s race against ChatGPT’s backers OpenAI and Microsoft Corp (MSFT.O), billions of dollars in investment and still untold advertising and cloud revenue from new AI programs are at stake.
Google’s warning also reflects what is becoming a security standard for businesses, which is to warn staff about using publicly available chat programs.
A growing number of companies around the world have set up guardrails on AI chatbots, including Samsung (005930.KS), Amazon.com (AMZN.O) and Deutsche Bank (DBKGn.DE), the companies told Reuters. Apple (AAPL.O), which has not returned requests for comment, reportedly has.
Some 43% of professionals used ChatGPT or other AI tools in January, often without telling their bosses, according to a survey of nearly 12,000 respondents, including from leading US companies, conducted by the networking site Fishbowl.
In February, Google told staff testing Bard before launch not to give it internal information, Insider reported. Now Google is rolling out Bard to over 180 countries and in 40 languages as a springboard for creativity, and its warnings extend to its code suggestions.
Google told Reuters it has had detailed discussions with Ireland’s data protection commission and is answering regulators’ questions, following a Politico report Tuesday that the company was delaying Bard’s EU launch this week pending more information on the chatbot’s impact. on privacy.
WORRY ABOUT SENSITIVE INFORMATION
Such technology can compose emails, documents and even software itself, promising to speed up tasks immensely. However, this content may contain misinformation, sensitive data, or even copyrighted passages from a “Harry Potter” novel.
A privacy statement from Google, updated June 1, also states, “Do not include confidential or sensitive information in your Bard conversations.”
Some companies have developed software to address such issues. For example, Cloudflare (NET.N), which defends websites against cyberattacks and offers other cloud services, is marketing an ability for businesses to tag and restrict certain data from flowing externally.
Google and Microsoft also offer enterprise customers conversational tools that come with a higher price tag but don’t include data in public AI models. The default setting in Bard and ChatGPT is to save users’ conversation history, which users can delete.
It “makes sense” that companies don’t want their staff to use public chatbots for work, says Yusuf Mehdi, Microsoft’s chief marketing officer.
“Companies are taking a pretty conservative stance,” says Mehdi, explaining how Microsoft’s free Bing chatbot compares to its enterprise software. “Our policy is much stricter there.”
Microsoft declined to comment on whether it has a blanket ban on personnel entering confidential information into public AI programs, including its own, though another executive there told Reuters that he had personally limited its use.
Cloudflare CEO Matthew Prince said typing confidential stuff into chatbots was like “unleashing a bunch of PhDs into all your private data.”
Reporting by Jeffrey Dastin and Anna Tong in San Francisco Edited by Kenneth Li and Nick Zieminski