Uber rolls out ChatGPT Work to 30,000 employees: 11 companies that have restricted AI tools
Uber is giving more than 30,000 employees access to ChatGPT Work, after first testing it with 1,000 people. The company is now making this AI tool available to teams around the world. It can help employees complete complicated tasks that involve several steps.
According to OpenAI's announcement, ChatGPT Work can connect to workplace apps such as Google Drive, Gmail, Google Calendar and Slack. Employees can use it to find information, prepare notes for meetings, study data, make presentations and complete tasks that involve several apps.
This decision shows an important change in how companies use AI at work. Businesses want to give employees more useful AI tools while also keeping private information safe. Several large companies have previously limited the use of ChatGPT and other public AI tools because they were worried about private company information, computer code and following rules.
Here are 11 companies that have limited the use of AI at work, according to reports from Fortune, Bloomberg, Axios, Forbes and HR Brew. These restrictions were reported in the past. They do not necessarily show what each company's rules are today.
1. Apple limited ChatGPT use to protect its private information
Apple works hard to keep its future products and software plans secret. Sharing private information with a public AI tool could put those secrets at risk.
According to reports from 2023, Apple limited employees' use of ChatGPT and other outside AI tools, including GitHub Copilot. The company was worried that employees might accidentally share private computer code or other confidential company information.
This decision showed a problem technology companies face: AI can help employees finish work faster, but companies must also protect the information employees enter into outside AI tools.
2. Samsung banned generative AI after employees shared sensitive information
Samsung Electronics banned generative AI tools on company devices and internal networks after employees reportedly shared sensitive information with ChatGPT, according to Bloomberg.
Employees reportedly entered private company computer code and other internal information into ChatGPT while asking it to help with their work. Samsung's response showed how using AI to make work easier can quickly create problems with keeping information safe.
Samsung became a well-known example of what can go wrong when employees use public AI tools without enough safety rules.
3. JPMorgan Chase limited ChatGPT because of concerns about following rules
For banks, it is not enough to know whether an AI tool can do a job. They also need to know whether employees can use it while following the bank's rules and legal requirements.
JPMorgan Chase limited employees' access to ChatGPT, according to reports from Axios and other news outlets in 2023. The decision was connected to rules about software from outside companies and concerns about protecting private information and meeting legal requirements.
The restriction showed the problems banks face when deciding whether employees should use public AI tools for work involving private business or customer information.
4. Bank of America put ChatGPT on its list of apps employees could not use
Bank of America included ChatGPT on a list of apps that employees were not allowed to use for work, Fortune reported in 2023.
This was part of the bank's wider rules for software made by outside companies. Banks can reduce the risk of private information reaching outside services by controlling which apps employees are allowed to use.
5. Citigroup limited employees' access to ChatGPT
Citigroup also limited access to ChatGPT through its rules for software made by outside companies, according to Fortune.
The report said the bank was studying both the possible benefits and risks of the technology. This showed how financial companies could learn about generative AI while still limiting employees' use of public AI tools.
6. Goldman Sachs limited public AI tools while looking at ways to use AI internally
Goldman Sachs limited employees' access to ChatGPT through its rules for outside software, Forbes reported in 2023.
At the same time, the investment bank was looking into ways to use generative AI inside the company. These included making short versions of longer information and helping employees with research. This showed the difference between blocking AI tools that a company cannot properly control and testing AI in a workplace where the company has more control.
7. Deutsche Bank blocked employees' access to ChatGPT
Deutsche Bank blocked employees from accessing ChatGPT, according to reports from Fortune and Forbes.
The restriction was linked to worries about private banking information being shared or leaked. The bank was also studying how AI could be used in systems where it could be carefully controlled.
8. Northrop Grumman limited public AI tools
For a defence company, sharing sensitive information could cause problems that go far beyond exposing ordinary business secrets.
Northrop Grumman limited access to public AI tools, according to HR Brew. Its approach showed why companies need to carefully check outside services before allowing employees to use them for work involving sensitive company or defence information.
9. Verizon blocked ChatGPT on company systems
Verizon made ChatGPT unavailable through its company systems, Fortune reported in 2023.
The telecommunications company was worried about losing control of sensitive information, including customer details and computer code. The restriction was meant to stop employees from using ChatGPT through workplace systems without proper controls.
10. Accenture limited generative AI use for client-related work
Consulting companies often work with private information belonging to many different clients. This means using public AI tools raises questions about keeping that information safe and handling client work responsibly.
Accenture introduced limits on employees using generative AI for writing computer code and doing client-related work, according to HR Brew. The reported guidance said employees needed proper permission for certain uses and warned them not to share private company or client information.
There is an important difference here. Limiting AI use for certain tasks or types of information does not always mean a company has banned every AI tool for all employees.
11. Amazon warned employees not to share private information with ChatGPT
Amazon warned employees not to enter private company information into ChatGPT, according to Fortune, which cited a Business Insider report.
The company also told engineers to use its own AI coding assistant, CodeWhisperer, in certain situations. According to the reports, Amazon's approach focused on protecting private information rather than completely banning every generative AI tool.
Uber's approach shows how companies are changing the way they use AI
Uber's decision helps explain the bigger debate about AI at work. Instead of only limiting AI use, Uber is giving more employees access to an AI tool designed for businesses. The tool can connect to approved workplace apps. The company is also supporting employees with training and an internal collection of resources for learning AI skills.
This comparison does not prove that Uber's approach is safer. It also does not mean the other companies still have the same restrictions they introduced earlier. It simply shows that companies have chosen different ways to let employees use AI while protecting sensitive information.
For businesses in India thinking about using AI tools designed for companies, there are some important practical questions. What information should employees be allowed to share? Which company systems should an AI tool be allowed to access? And how will the company check and control what the AI does?
As AI tools move beyond simply answering questions and start completing tasks with several steps, companies will need rules that cover more than just who can use AI. They will also need to decide what information and systems AI can access and what actions it is allowed to take.