Blog - Tag: AI
AI ‘Botsitting’ Is Costing Employees Hours
AI ‘Botsitting’ Is Costing Employees Hours
Artificial intelligence is supposed to save employees time by taking mundane tasks off their plates and helping them work more efficiently. But a new study suggests workers are spending nearly as much time supervising AI as they spend getting useful work out of it.
A report from workplace AI company Glean found that employees spend 37% of their AI-related time “botsitting” — checking AI’s output, correcting errors, rerunning prompts and providing missing context. That compares with 36% of their AI time spent using the technology to produce work.
In practical terms, for every hour employees spend getting usable work from AI, they spend roughly another hour dealing with its shortcomings, the study concludes. For employers investing heavily in AI, that hidden labor may complicate the technology’s promised return on investment.
Employers may need to reconsider how extensively they integrate AI into their operations if employees spend an inordinate amount of time achieving results similar to those they produced without it.
The human cost of AI
Glean defines botsitting as the work required to make AI usable, including supplying missing context, reviewing its output, debugging mistakes, rerunning prompts and cleaning up answers that sound convincing but are wrong.
The company surveyed 6,000 digital workers and found that 87% use AI at work. About three-quarters said the technology makes them more productive, estimating that they save about 11 hours a week.
However, just 13% said AI has improved their organization’s overall performance. One possible reason for this disconnect is the amount of human oversight the technology requires. According to the report:
- Workers spend an average of 6.4 hours a week botsitting.
- Employees spend more than a quarter of their AI-related time learning how to use the technology and building AI agents.
Employers should therefore be cautious about counting every hour theoretically saved through AI as a genuine productivity gain. If an employee uses AI to complete a task faster but then has to verify facts, correct errors and repeatedly refine the output, some of the time saved is lost.
An invisible workload
Botsitting can also create an HR challenge because much of the work may go unnoticed.
Employees are often expected to incorporate AI into their jobs without having other responsibilities reduced. As a result, reviewing AI-generated output can become an added responsibility rather than a replacement for other work.
Glean found that employees who spend more time refining AI output are more likely to report feeling exhausted by the process. Workers who spend the most time botsitting are also significantly more likely to be looking for another job.
For managers and HR departments, these findings raise questions about workload, burnout and whether employees are receiving enough training and support.
Making AI work for employees
Employers should consider the human labor required when calculating the costs and benefits of AI. Simply purchasing AI tools and encouraging employees to use them may not deliver the expected productivity gains.
Organizations can start by identifying which tasks AI handles reliably and which require substantial human review. Managers should also track how much time employees spend correcting AI output rather than assume it is saving time.
Training is just as important. Employees need guidance on which tools to use, how to write effective prompts, what information can safely be entered into AI systems and when AI output requires additional verification.
Most importantly, businesses may want to view AI as a tool that supports employees rather than one that operates independently. Human judgment remains essential.
Study Finds AI Hiring Tools May Increase Bias Risks
A new study has found that artificial intelligence AI hiring tools can disproportionately screen out Black and Asian job applicants and repeatedly reject the same candidates across multiple employers using the same technology.
The findings raise new concerns for employers as AI increasingly becomes a fixture in recruiting. According to the World Economic Forum, about 90% of firms now use artificial intelligence to help vet job applicants.
Researchers at Stanford’s Institute for Human-Centered Artificial Intelligence analyzed 4 million job applications submitted to more than 150 employers using the same third-party hiring platform. They found that 26% of Black applicants and 15% of Asian applicants applied for positions where the AI system produced outcomes that met the Equal Employment Opportunity Commission’s standard for adverse impact. That means about 40,000 applications by these applicants could have advanced to the next stage of the hiring process.
The study’s findings mean that employers need to be mindful of potential discriminatory results from using AI during the hiring process. Courts have historically held employers liable for practices that create disparate impact on protected groups, and several recent lawsuits have challenged the use of AI in hiring under federal anti-discrimination laws.
Bias may be hidden
The study found that discriminatory outcomes can be difficult to spot because aggregate data often mask problems.
For example, an AI system might recommend Black applicants for one type of job but reject them for another. When all positions are combined, the disparities may disappear statistically even though significant differences exist for individual jobs.
That distinction matters because courts typically analyze disparate-impact claims on a position-by-position basis.
Perhaps the study’s most significant finding involves what researchers call “algorithmic monoculture.”
Many employers rely on the same small group of third-party vendors to screen candidates. As a result, the same algorithms may influence hiring decisions across hundreds of companies.
The researchers found that applicants who applied to multiple jobs screened by the same AI platform were more likely to be rejected by every employer than would be expected if each company made decisions independently. One in 10 applicants who submitted four applications through the platform was rejected by all four employers.
The takeaway
The study reinforces a key legal truth: employers remain responsible for the tools they use.
Even when using systems from third-party vendors, courts are likely to hold employers accountable for discriminatory outcomes. As a result, employers should treat AI governance as a central part of their compliance efforts.
Employers that use AI in recruiting may want to consider the following safeguards:
- Conduct adverse-impact analyses by individual job position.
- Require vendors to provide validation studies and bias-testing data.
- Maintain human review of applicants screened out by AI.
- Periodically audit hiring outcomes for potential disparities.
- Document how AI systems are selected, tested and monitored.
- Establish oversight teams that include HR, legal and technology personnel.
- Train managers on the limitations and risks of automated decision-making.
- Closely monitor evolving federal and state regulations governing AI in employment.
Insurers Start Excluding AI Risk in Commercial General Liability Policies, More
Some insurers have begun introducing exclusions for artificial intelligence-related claims from standard business insurance policies, creating potential coverage gaps for businesses that rely on AI tools for marketing, customer service, product development or daily operations.
The changes come after the Insurance Services Office, the industry’s clearinghouse for policy language, introduced three new artificial intelligence exclusions for commercial general liability policies that insurers are beginning to add to coverage forms.
Roughly 86% of all U.S. property/casualty insurance policies contain some form of ISO language, meaning these exclusions could soon become widespread and leave coverage gaps for many employers when their CGL policies come up for renewal. Insurers are also starting to add similar language to other policies with a liability component.
New coverage gap
The three new ISO endorsements include:
- CG 40 47 — The broadest form, excluding coverage for bodily injury, property damage or personal/advertising injury arising out of generative AI.
- CG 40 48 — A narrower endorsement excluding only personal and advertising injury claims tied to AI.
- CG 35 08 — An exclusion applying to products and completed operations liability coverage.
These endorsements could affect how coverage applies to certain AI-related claims, depending on policy language and endorsements.
One of the largest concerns involves Coverage B of the CGL policy, which traditionally covers claims such as defamation, invasion of privacy, misappropriation of advertising ideas or certain intellectual property-adjacent disputes. Under the new exclusions, those claims may no longer be covered if they arise from AI-generated text, images, audio, video or code.
Even businesses using third-party AI tools, rather than developing their own systems, may still trigger the exclusions. In some cases, incidental use of AI may be enough.
The businesses likely to feel the greatest impact are those integrating generative AI deeply into operations, including:
- Marketing and advertising firms using AI-generated campaigns,
- Technology companies embedding AI into products or software,
- Manufacturers relying on AI-assisted product design,
- Professional service firms using AI to draft documents or communications,
- Retailers deploying AI chatbots or recommendation engines,
- Nonprofits using AI for outreach or donor engagement, and
- Employers using AI tools in hiring or HR decisions.
Insurers are not stopping with general liability coverage. AI exclusions are also beginning to appear in:
- Directors and officers liability,
- Employment practices liability,
- Fiduciary liability,
- Cyber, and
- Errors and omissions policies.
Some insurers have already received regulatory approval for AI exclusions in Florida, Connecticut and Maryland. Others, including W.R. Berkley, have adopted broader exclusions that eliminate coverage for claims arising out of the use, deployment or development of AI across multiple lines of coverage.
What you can do
Businesses should expect insurers to ask more detailed questions about AI usage during renewals and underwriting. Companies that fail to evaluate potential coverage gaps could find themselves uninsured for lawsuits, regulatory investigations or shareholder claims tied to AI-generated content or decision-making.
Organizations should consider taking the following steps:
- Identify where AI is being used throughout the organization.
- Strengthen internal AI governance and oversight procedures.
- Require human review of AI-generated content and decisions.
- Train employees on acceptable AI use.
- Evaluate contracts with AI vendors and third-party providers.
- Discuss AI exposures and coverage gaps with us before renewal.
- Explore specialized protection options.
Some organizations may ultimately need dedicated technology errors and omissions coverage, cyber liability insurance or emerging standalone AI insurance products designed to address AI-related risks. Call us with questions.
New AI-in-Hiring Rules Are in Effect: What You Need to Know
Starting Oct. 1, 2025, any California employer that uses artificial intelligence and other automated tools in recruiting, hiring, promotion and related human resources decisions will have to ensure that the tools don’t discriminate against protected classes.
The new regulations, promulgated by California’s Civil Rights Department, cover any “automated decision system” (ADS) which the rules broadly define to include any computer-based process that makes or influences employment decisions, such as:
- Artificial intelligence,
- Machine learning,
- Algorithms,
- Statistics, and
- Other data-processing techniques.
If your firm uses AI or another data-driven system in hiring, you’ll want to beef up record-keeping and set testing procedures to ensure that the tools you use comply with the new regulations.
What counts as an “automated-decision system”
Examples of systems that are covered by the new regulations include:
- Résumé screeners — These may favor applicants who use certain wording, which can disadvantage older workers or those from different cultural or educational backgrounds.
- Targeted job-ad delivery — Tools may push job ads to preferred genders, age groups, races and other protected classes.
- Puzzle or game-style assessments — These tools may screen out people with certain physical or neurological conditions.
- Voice and facial analysis tools — Tools that assess “enthusiasm” or “communication style” may produce biased results against applicants with disabilities, speech differences or accents.
Basics of the new rules
Discrimination risk — It is unlawful to use an ADS or other selection criteria that discriminate based on any protected characteristic such as race, gender and ethnicity. Crucially, an employer can be liable even without intent if the ADS causes an adverse disparate impact on a protected class.
Anti-bias testing — Employers are required to perform anti-bias testing of their automated systems. In any investigation or lawsuit, regulators and courts may look at six factors to determine whether an employer took reasonable steps to avoid discrimination:
- Quality of the testing
- Efficacy (how well it detects bias)
- Recency (how current it is)
- Scope (which systems or data were tested)
- Results of the testing or due diligence
- The employer’s response to those results (what was changed or fixed afterward)
Failing to conduct or document bias testing could weigh against an employer in a discrimination case.
Record-keeping — The rule requires employers to keep ADS-related records for four years.
What you can do
If you use an ADS system in your personnel decisions, focus on the following to comply with the new rules:
Tracking — Track your ADS system’s involvement in recruiting, hiring, promotion, training selection, performance screens or advertising. Include vendor tools and “off-the-shelf” filters.
Testing — Build a defensible bias-testing program and document the six factors that regulators will look at:
- Quality,
- Efficacy,
- Recency,
- Scope,
- Results, and
- Your response.
Planning — Establish a plan to regularly test your ADS systems for bias-tainted decisions. Most importantly, if you detect deficiencies, document the steps you took to address the problems.
The takeaway
One of the keys to a successful defense is showing you have taken steps to remedy issues with tools that you use in employment decisions. That means being able to show that you have ensured your data-driven personnel tools do not discriminate.
As a side note, employers should expect more AI-related legislation in the years to come as more companies use it in their day-to-day operations.