Unpacking AI: Opportunities, Risks and Insurance Considerations for Businesses
Generative artificial intelligence (AI) programs like Open AI’s ChatGPT and Dall-E thrust AI into overnight stardom.
Generative AI is a processing tool built on deep machine learning that can digest colossal amounts of data. It can synthesize, translate, analyze and summarize dense information in a human tone. Its responses and interactions may seem human, but it’s simply programmed that way.
It’s already reimagining uses for think tanks and impromptu ideas naturally occurring on workplace productivity platforms. For example, you could use ChatGPT to parse posts from your engineering department’s collaboration platform to create short-form creative brainstorm reports. You could then share these posts with your C-suite.
But for all the perks of AI, there are just as many risks. As the dust from the AI media storm settles, practical questions remain: How can AI help your business, what are the risks and how do you insure against them?
AI, a new “old” technology
Conceptualized in the 1940s during the dawn of cybernetics, the makings of today’s AI have been around for a while. AI took center stage as big data and computer processing capabilities increased in the 2000s.
From smart homes to health care analysis tools, AI has been changing how we relate to computers. AI is transforming the human experience, including things like:
- Data analysis and benchmarking
- Assistive services
- Health care services
- Communication services
- Art, writing and music interaction
- Training and education initiatives
- Recruiting and people management
- Predictive modeling
- Project management
- Critical thinking and strategic development
AI implementation considerations
As with any new technology, you might face challenges involving data management, initial setup and staff training. Start by assessing how AI could benefit your business and if you have the budget to onboarding it, customizing it and training your employees on how to use it. An unbiased consultant could help you decide if AI is right for your business operations.
The White House published a Blueprint for an AI Bill of Rights for employers considering AI. It includes considerations from multiple perspectives:
- Harmful outcomes involving AI, like algorithmic bias, hiring discrimination and patient care
- Helpful effects of applying AI, like improved crop yields in farming and predictive storm path modeling
- Human concerns like customer experience, consumer choice and privacy
While it’s true that AI can be resource-intensive at the outset, it could offer long-term advantages if you appropriately implement it. Implementation requires a lot of planning, including addressing cybersecurity, intellectual property and data privacy concerns. And using AI could impact job roles.
Make a plan to navigate these changes, including adopting policies around AI and being transparent with how you intend to use AI in your operations.
Examples of AI in business
Here are some real-life examples of ways businesses are training and using AI.
Adding closed captioning to eyewear
An eyeglass manufacturer is embedding a form of augmented reality into its frames and tethering it to an AI-powered smartphone app. The AI recognizes voices and transcribes spoken conversations onto the inside of the frames so the wearer can read the chat in real time.
Taking restaurant orders
An independently owned restaurant is using an AI chatbot to take online orders, providing real-time answers to customer inquiries about menus, ingredients and availability. The AI chatbot has freed restaurant staff to handle more complex customer inquiries and orders.
A fashion retail company is using a predictive AI tool to manage inventory and sales. After initial failures leading to massive overstock, the retailer retrained its AI to include multiple inputs, from weather sites to social media. This honed the AI tool’s knowledge, making it more effective at predicting trends and anticipating inventory.
All of these AI applications required training. As the last example illustrates, planning and sourcing reliable data are key. And in any case, you can expect a lot of trial and error.
But as AI becomes more prevalent, the risk exposures will be evident across the business landscape. Unfortunately, threat actors are a part of that landscape, too.
AI risk exposures and ways to insure them
Since AI is still rapidly evolving, insurance for businesses using generative AI is a relatively new and complex field.
There isn’t a specific insurance for AI liability yet. Keep in mind that the liability risks will evolve as AI evolves. Connect with your independent insurance agent and take advantage of annual coverage reviews. Let them know if you’ve changed your operations in any way.
You can insure against your risks using layers of insurance. You might already have some of these in place:
Cyber liability or data breach insurance covers legal fees and expenses associated with cybercrime. For example, if a cyberattack exposes personal data, cyber liability will help pay for network restoration and required personal identity monitoring. Any business that uses an internet-connected computer or device needs this insurance.
Technology errors and omissions (E&O) insurance covers software developers and technology consultants for negligence, mistakes or failures to perform their professional duties. If you recommend AI products or work in the technology field, this coverage is for you.
Intellectual property insurance covers litigation expenses for lawsuits involving infringement of intellectual property. This is essential coverage since generative AI can output content that infringes on others’ intellectual property.
Media liability insurance covers allegations that your content harmed someone financially or caused them emotional distress. Claims could involve things like copyright infringement or defamation. Reporters, authors, social media influencers, advertisers and marketers typically have this coverage. You should consider a policy if you’re using AI chatbots to generate content online or interact with the public. An AI program could malfunction or get hacked and generate harmful content, which you’d be liable for.
General liability insurance covers legal fees and payments resulting from physical harm. For example, if a robot or AI-powered machine you operate causes an accident, your general liability insurance will respond.
Directors and officers (D&O) insurance covers your management team if you get sued. D&O coverage helps with legal fees and settlement costs. D&O coverage is essential if you’re a board member outside of your business. Board members, directors and officers can be held personally liable for injuries and damages.
Professional liability insurance can help protect your business against claims of inadequate work or negligent actions. It’s also known as E&O insurance.
Employment practices liability insurance (EPLI) helps protect you when an employee sues you. General liability won’t cover you for lawsuits related to employee complaints. General liability covers third-party claims, like ones made by customers or the public. EPLI covers you when an employee sues you for things like:
- Sexual harassment
- Wrongful termination
- Breach of employment contract
- Negligent evaluation
- Failure to employ (hire) or promote
- Wrongful discipline and deprivation of career opportunity
- Wrongful infliction of emotional distress
- Mismanagement of employee benefit plans
For example, if you launch an AI candidate screening program without robust supervision and testing, you’re at risk. Companies use AI to sift resumes, create knowledge and skill assessments, check references and conduct interviews with face and voice analysis.
Unchecked, any of these techniques could result in a failure to employ lawsuit. Failure to employ is when job candidates feel they weren’t selected because of a legally protected characteristic (such as religion, age, ethnicity, gender, disability, skin color, sexual orientation or race).
A word on vendor risk exposures
Even if you don’t directly use AI, vendors and other parts of your supply chain might. Take time to understand how they’re using AI and their ability and willingness to take responsibility for AI faults or issues that may arise.
For example, say one of your vendors uses AI to place orders and send confirmations. The AI malfunctions and fails to place incoming orders, but sends confirmations anyway. The resulting backlog creates a supply chain bottleneck. Will your vendor handle the liability?
Be clear on the division of liability.
Sometimes AI risk can come directly from the AI vendor. Follow the same best practices you would with any technology and use a reputable tech company. Do they offer employee training and ongoing support? Research their warranties and ask for proof of their insurance, like technology E&O coverage. Be wary of companies that absolve themselves from all liability after implementation.
Consult your lawyer before entering into any contracts or agreements.
Cyber liability insurance challenges
You’ll need to insure the challenges associated with using AI. Use a seasoned independent agent who understands the risks associated with AI generally and within your industry. Be prepared to discuss how you’re using AI. Your agent will explain risk exposure areas and techniques to correct the risks.
You might have more cyber liability than you realize. For example, if you don’t have a cyber liability or risk response policy, you may need to correct your risks and make a plan to qualify for insurance. Implementing strong cybersecurity measures is a prudent move, and it could improve your insurance rates.
It’s also important to note that laws regulating AI use vary by country. If you use AI across international locations, ensure your insurance and usage policies comply with local laws and regulations.
Bolstering your cyber liability application
Most insurance companies will request information to assess your vulnerability to an attack. They’ll also look at:
- Your data-sharing, collection and storage practices
- Your cybersecurity strategy
- Your cyber incident response plan
Since AI is a computer tool, it straddles many of these larger risks. The insurance company might ask you direct questions about your AI usage, but often, cyber liability applications involve listing the software you regularly use. They’d be able to infer that it’s AI. Be truthful about your computer information and cybersecurity protocols. If you omit or embellish a process, you risk a claim denial in the future. You’re better off taking the hit on higher premiums than being denied assistance when you need it most.
It can be intimidating to put your cybersecurity under scrutiny. But having reliable cyber liability insurance can mean the difference between bankruptcy and recovery after a cyberattack. The Cybersecurity & Infrastructure Security Agency has information to help business owners evaluate their risks and find solutions. Get a cybersecurity and risk audit from a reputable technology firm.
Stay vigilant about protecting against liability risks
As AI continues to ramp up, it will affect every industry. Treat AI like you would any new computer tool. Take steps to secure it from cyberattacks and stay vigilant about creating and updating policies for its uses within your business.
Contact us if you’re launching an AI-based initiative at your company. They can help you find insurance solutions to protect what matters most.