How Companies Can Make Sure AI Respects Civil Rights

Companies must move carefully and in a principled way when building and using AI. Failing to respect civil rights – from privacy to free speech to anti-discrimination laws – will land your company in the hot seat, with expensive lawsuits, government investigations and fines, bad press, and loss of customer trust. This is especially true in the current moment, as laws are rapidly changing with many state lawmakers introducing bills related to AI.

Privacy & Free Speech: It’s Good for Business, our ACLU Business Primer, includes more than 150+ case studies and tips to bake in the right protections from the start and avoid problems down the line.

• Avoid facing huge government fines like Meta did for using people’s personal data to train AI models
• Don’t be forced like Everalbum to delete your algorithms because they were improperly built on people’s personal information.
• Avoid expensive lawsuits, like those facing Clearview and OpenAI for harvesting people’s information from the internet.
• Don’t let your reputation take a hit and be dinged publicly like Zoom or LinkedIn for trying to quietly make changes to use people’s information for AI.

You’ll want to tread carefully and check out the dozens of real-life business case studies and additional resources about why and how to bake in the right protections right from the start and avoid problems down the line in our ACLU Privacy and Free Speech Primer.

1. RESPECT YOUR DATA
   Protecting your users’ privacy in the age of AI requires you to be thoughtful about the data you collect and hold. By carefully considering the costs and benefits of collecting data and by properly safeguarding the information that you do collect, you may prevent privacy harms and increase consumer trust in your product.
2. PLAN AHEAD
   AI raises many new questions and problems involving data collection and storage. Save time, money, and even your reputation by maintaining privacy and security practices that are holistic, regularly re-evaluated, and prepared for potential data security issues and legal demands.
3. BE TRANSPARENT
   The first step in establishing and maintaining a trust-based relationship with your users is giving them the information they need to make informed decisions and understand the AI product they’re using. Doing so not only helps prevent surprises that can lead to backlash, it can also build loyalty among your current users and help you recruit new ones.
4. PARTNER WITH YOUR USERS
   Even if you plan to offer your product “for free” and generate revenue from advertising or other means, it is in your best interest to treat your users as partners: recognizing and respecting their expectations, giving them the tools to make their own decisions about their personal information, and standing up for them when they are unable to defend themselves. By doing so, you may not only avoid the consequences when users are unpleasantly surprised, you may find that users who trust you are more willing to pay for or engage with your service.
5. ENCOURAGE USERS TO SPEAK FREELY
   AI should not threaten people’s ability to speak and express themselves. The more freedom your users have to express themselves freely, the more likely they are to interact deeply with your service, with lasting benefits to everyone involved.
6. MODERATE CAUTIOUSLY
   AI raises many thorny questions around moderation. Any policies and mechanisms must be regularly evaluated to address civil liberties and civil rights impacts on users, particularly users whose voices have historically been marginalized.
7. PROMOTE CREATIVITY
   With AI, as with other products, encouraging your customers to use your content or service in new and innovative ways may attract more paying users, while limiting their ability to enjoy your service could drive them to less restrictive competitors.
8. SPEAK UP FOR FREE SPEECH
   To gain the respect of users as a champion for their free speech, you need to do more than just allow users to express themselves on your platform or service – you need to proactively protect them. With so much uncertainty around AI, earning a reputation as a defender of your users’ rights can be a valuable way to build trust with your current users and attract new ones.