Alphabet Inc, the parent company of Google, is urging its employees to exercise caution when using AI chatbots, including its own Bard program, amid concerns about the security of confidential information, according to sources familiar with the matter. The company has instructed its employees not to enter confidential materials into AI chatbots, citing a long-standing policy to safeguard information.
Chatbots such as Bard and ChatGPT are advanced programs that utilize generative artificial intelligence to engage in conversations with users and provide responses to various prompts. Human reviewers may review these chat interactions, and researchers have discovered that similar AI models can reproduce the absorbed data during training, posing a potential risk of data leakage.
Alphabet has also advised its engineers to avoid directly using computer code generated by chatbots, as it may carry certain risks. The company confirmed that Bard may make undesired code suggestions, but it still assists programmers. Google emphasized its commitment to transparency regarding the limitations of its technology.
These concerns reflect Google’s efforts to mitigate potential business harm arising from the competition with ChatGPT, developed by OpenAI and supported by Microsoft Corp. The race between Google and its competitors in the AI field holds substantial investment opportunities and significant revenue potential from new AI programs, including advertising and cloud services.
Google’s cautious approach aligns with a security standard increasingly adopted by corporations worldwide, which involves warning personnel about the usage of publicly-available chat programs. Several companies, including Samsung, Amazon.com, and Deutsche Bank, have implemented similar guardrails on AI chatbots. While Apple did not provide a comment, it is reportedly also taking precautions in this regard.
A survey conducted by the networking site Fishbowl revealed that approximately 43% of professionals were using AI tools like ChatGPT as of January, often without informing their superiors. In February, Google instructed its staff testing Bard not to share internal information, and the company is now expanding Bard’s availability to over 180 countries and 40 languages, promoting it as a platform for creativity. The cautionary measures extend to the code suggestions provided by the chatbots.
Google has engaged in detailed discussions with Ireland’s Data Protection Commission and is addressing regulators’ questions regarding Bard’s impact on privacy. This follows a Politico report suggesting that Google had postponed the launch of Bard in the European Union pending further information.
The use of AI chatbots poses concerns regarding the inclusion of sensitive information in their generated content. While the technology offers the promise of accelerating tasks like drafting emails, documents, or software, it also runs the risk of incorporating misinformation, sensitive data, or copyrighted material. Google’s updated privacy notice, effective from June 1, explicitly advises users not to include confidential or sensitive information in their Bard conversations.
To address these concerns, companies have developed software solutions. Cloudflare, for instance, offers a capability that allows businesses to tag and restrict certain data from flowing externally. Google and Microsoft also provide conversational tools to enterprise customers, with enhanced security measures that prevent data absorption into public AI models. By default, Bard and ChatGPT save users’ conversation history, although users have the option to delete it.
Yusuf Mehdi, Microsoft’s consumer chief marketing officer, remarked that it is understandable for companies to discourage their employees from using public chatbots for work purposes. He emphasized the stricter policies in place for Microsoft’s enterprise software compared to its free Bing chatbot. While Microsoft did not comment on whether it has a blanket ban on employees entering confidential information into public AI programs, another executive indicated personal restrictions on their usage.
Matthew Prince, CEO of Cloudflare, likened inputting confidential matters into chatbots to allowing a group of Ph.D. students access to private records, emphasizing the potential risks associated with it.
The potential risks associated with entering confidential information into chatbots are significant. Companies must be vigilant about protecting sensitive data from unauthorized access or exposure. The inclusion of proprietary or copyrighted material in AI-generated content could result in legal consequences and intellectual property disputes. Additionally, the dissemination of misinformation through chatbots can have detrimental effects, leading to reputational damage or the spread of false information.
Recognizing these risks, some companies have taken steps to address the concerns surrounding AI chatbots. Cloudflare, for example, offers a solution that allows businesses to tag and restrict certain data from being transmitted externally, providing an additional layer of data protection. Google and Microsoft also offer conversational tools with enhanced security measures for enterprise customers, ensuring that sensitive information remains within controlled environments.
The cautionary approach adopted by Alphabet Inc and other companies in the industry reflects a growing consensus on the need for strict guidelines and policies surrounding the use of AI chatbots. While these chatbots can provide convenience and efficiency, there is a need to balance their benefits with the potential risks they pose to data privacy, security, and intellectual property rights.
As the use of AI chatbots becomes more prevalent across industries, organizations must establish clear guidelines for employees regarding their usage. Companies should educate their staff about the risks associated with entering confidential information into public chat programs and encourage them to exercise caution when interacting with AI chatbots. Regular training programs and awareness campaigns can help employees make informed decisions and understand the importance of safeguarding sensitive data.
Moreover, regulatory bodies and privacy advocates play a vital role in ensuring that companies adhere to best practices when it comes to AI chatbots. Close collaboration between technology companies and regulators can lead to the development of comprehensive frameworks and standards that protect user privacy and prevent data breaches.
