AI Training as a Side Job: Equipment, Schedules, Burnout and Ethics

Can I do annotation work using only an iPad or iPhone?

Only for some work. TELUS says a computer or smartphone can be enough for certain projects, but much evaluation work needs a full desktop browser, and Mercor's required Insightful time tracker does not run on iPads or phones. Plan on a laptop or desktop for most platforms.

Source: TELUS Digital AI Community: FAQ; Mercor Docs: Use Insightful for Time Tracking

What hardware do I need before taking the assessment?

A Windows or Mac computer, a current Chrome browser, and a stable internet connection cover most platforms. AI interviews add a working camera and microphone; Mercor recommends headphones and an external microphone. TELUS says some projects need a webcam, microphone or specific apps, stated in the project details.

Source: Mercor Docs: AI Interview; TELUS Digital AI Community: FAQ

Will a Chromebook work for DataAnnotation projects?

DataAnnotation does not publish device requirements, and its FAQ asks only for reliable internet. Browser-based tasks may work on a Chromebook, but platforms that need desktop tracking apps do not support them; Mercor, for example, says Chromebooks cannot run its time tracker. A Windows or Mac laptop avoids the risk.

Source: DataAnnotation: FAQs; Mercor Docs: Use Insightful for Time Tracking

Is a cheap backup laptop adequate for annotation while traveling?

Usually, for browser-based work: annotation needs a current browser and stable internet more than processing power. Check any required desktop app's operating system minimums (Mercor's tracker needs Windows 10 or macOS 11 or later), and remember travel can raise location and VPN issues.

Source: Mercor Docs: Use Insightful for Time Tracking; Mercor Docs: Account Security and Conduct

Can two household members work on DataAnnotation using the same Wi-Fi?

DataAnnotation's FAQ does not address it. Each person needs their own account, own identity verification and own payment account, and must do their own work. Other platforms are stricter: TELUS says some programmes accept only one member per household. Ask support before both applying.

Source: DataAnnotation: FAQs; TELUS Digital AI Community: FAQ

Could my family member's performance affect my account?

It should not if your accounts, payment methods and work are fully separate. The risk is linkage: OneForma prohibits using a payment account already linked to another user, and shared payment details or devices can look like one person running two accounts. Keep everything separate.

Source: OneForma: Freelancer Terms and Conditions

Can I refer someone who shares my internet connection?

Referral rules differ, and some programmes restrict households. TELUS limits certain programmes to one member per household, and Mercor bans self-referral and multiple accounts. Check the programme's rules, and make sure the other person applies, verifies and works independently.

Source: TELUS Digital AI Community: FAQ

What microphone is suitable for related paid audio work?

A USB microphone or quality headset in a quiet room is usually enough for interviews and many voice tasks. Audio data projects state their own requirements, sometimes strict ones, so read the project specification before buying anything; real platforms never require you to buy equipment through them.

Source: TELUS Digital AI Community: FAQ; Mindrift: How Mindrift pays contributors

Is annotation work manageable with fluctuating energy from a chronic illness?

Often, because many platforms let you choose when and how much to work: DataAnnotation and Alignerr set no minimum hours. Prefer platforms without weekly minimums, work in short focused sessions, and avoid claiming tasks with tight expiry timers on low-energy days.

Source: DataAnnotation: FAQs; Alignerr: FAQs

Which annotation jobs fit around a full-time job?

Platforms with no minimum hours and work available around the clock: DataAnnotation says projects run 24/7/365 with no minimum hours or daily logins, and Alignerr and TELUS both say the work can sit alongside another job. Avoid programmes with weekly minimums you cannot meet.

Source: DataAnnotation: FAQs; Alignerr: FAQs; TELUS Digital AI Community: FAQ

Are there platforms without mandatory hours or daily quotas?

Yes. DataAnnotation says you can work as little or as much as you want, with no minimum hours or daily login requirements. Alignerr says there are no minimum hour requirements. TELUS's AI Community programme is the exception, with a 10-hour weekly minimum.

Source: DataAnnotation: FAQs; Alignerr: FAQs; TELUS Digital AI Community: FAQ

Can I do this work only in evenings and weekends?

On flexible platforms, yes. DataAnnotation's work is available 24/7/365 and you choose when to work. Some projects have deadlines or time-zone-specific meetings, so check each project's expectations.

Source: DataAnnotation: FAQs

What should I do if two platforms assign me the same client project?

Raise it with one or both platforms before doing any work. Working the same client project through two vendors can breach exclusivity or confidentiality terms, and TELUS notes some projects have exclusivity clauses. Get written guidance.

Source: TELUS Digital AI Community: FAQ

Should I notify both platforms or drop one assignment?

Notify them and ask. If the terms of either prohibit overlapping work, drop the one that conflicts. Being open is safer than risking both accounts.

Source: TELUS Digital AI Community: FAQ

Am I allowed to work on other AI training platforms while doing DataAnnotation?

DataAnnotation describes its work as fully flexible and its FAQ sets no exclusivity rule, but check your agreement. Other platforms, such as Alignerr and TELUS, say the work can run alongside other jobs, while some TELUS projects carry exclusivity clauses stated in the job description.

Source: DataAnnotation: FAQs; Alignerr: FAQs; TELUS Digital AI Community: FAQ

Do different platforms depend on the same underlying AI clients?

Often, yes. A small number of AI labs buy training data from several vendors, so demand can rise and fall across platforms together, and two platforms can staff the same client. That is also why dry spells can cluster.

Can working very long hours damage quality and jeopardize access?

Yes. Evaluation quality drops with fatigue, and quality determines access: DataAnnotation says subpar contributions can lose future projects. Caps exist partly for this reason; Mercor says it limits weekly hours to support quality and avoid burnout.

Source: DataAnnotation: FAQs; Mercor Docs: Time Tracking and Pay Policies

Do I need to work every day while projects are available?

Not on flexible platforms; DataAnnotation has no daily login requirement. Projects with deadlines or limited batches reward checking in regularly, but you choose your days.

Source: DataAnnotation: FAQs

How much time off do experienced annotators take?

There is no standard, since you set your own schedule. Programmes with commitments set limits: TELUS lets AI Community members block out up to 30 days, with extensions in exceptional cases, and warns prolonged inactivity can remove access.

Source: TELUS Digital AI Community: FAQ

Can I work in short sessions when instructions alone take much of my available time?

Yes, by separating learning from doing: read the guidelines in one session, then do tasks in later short sessions with your own summary to hand. Avoid claiming tasks with expiry timers you cannot finish in a short session.

Why do tasks feel harder after a long period without work?

Guidelines fade from memory and projects often update them in the meantime. Re-read the current guidelines before your first task back, and expect to be slower for the first few.

Why do I burn out after a few annotation hours despite handling full-time work elsewhere?

Annotation is unusually dense: constant judgment, little variety and no social breaks. Two or three hours of focused evaluation can tire you more than a varied day job. Shorter blocks with real breaks help.

Source: Mercor Docs: Wellness Resources

How can I keep concentrating after two or three hours of annotation?

Work in timed blocks (for example 45 to 50 minutes) with a short break away from the screen, switch to a different task type between blocks, and stop when your accuracy drops. Mercor publishes optional wellness resources with quick resets for focus.

Source: Mercor Docs: Wellness Resources

How do I stop feeling guilty about taking breaks when paid tasks are available?

Remember that tired work costs more than it earns: low-quality submissions can lose you access to future projects, as DataAnnotation warns. Breaks protect your income. Set a schedule in advance so stopping is a plan, not a decision each time.

Source: DataAnnotation: FAQs

What break schedule helps maintain motivation and productivity?

A fixed rhythm works for most people: focused blocks of 45 to 60 minutes, a 10-minute break, and a longer break every few hours. Stop for the day when you notice repeated errors. Adjust to what keeps your quality steady.

How do full-time annotators avoid burnout?

By treating it like a job: set hours, a dedicated workspace, regular breaks, varied task types, and days off. Mercor, for example, offers wellness resources and notes support for people working with sensitive content.

Source: Mercor Docs: Wellness Resources

Does working from a cafe or changing environment help concentration?

It can for some people, but check the rules: public Wi-Fi and VPNs can trigger security flags on some platforms, and confidential project material should not be visible to others. A quiet change of room at home is safer.

Source: Mercor Docs: Account Security and Conduct

Should I follow a conventional workday routine?

If it helps you stay consistent, yes. Fixed start and stop times reduce decision fatigue and make it easier to protect rest. Flexibility is still useful for fitting around other commitments.

How do people sustain more than 20 hours of annotation per week?

Spread it across the week rather than long days, mix task types, take real breaks, and keep weekends or days off. Alignerr says many of its contributors work 20 to 25 hours a week, which is sustainable for many people when paced.

Source: Alignerr: FAQs

Does switching project types help or worsen mental fatigue?

Switching between blocks helps many people, because variety reduces monotony. Switching within a block can hurt accuracy, since each project has different rules. Change type at the break, not mid-session.

Will choosing easier projects or taking a break reduce future access?

Taking breaks generally does not on flexible platforms, though prolonged inactivity can on some; TELUS, for example, warns it may remove program access. Quality matters more than the difficulty of projects you choose.

Source: TELUS Digital AI Community: FAQ; DataAnnotation: FAQs

How do people reconcile AI training work with concerns about generative AI?

Common approaches are focusing on safety and accuracy work, which makes models less harmful, choosing projects that align with your values, and declining ones that do not. Many see shaping AI with expert judgment as better than leaving it unshaped.

Can improving model safety justify concerns about environmental or social harm?

It is a personal judgment. Safety and accuracy work reduces some harms from AI systems, but it does not address their energy use or labour effects. Weigh what the specific project does, and decline work you are not comfortable with.

Will improving AI eventually eliminate demand for human annotators?

For simple labeling, demand is already shrinking as models automate it. For expert evaluation, demand has grown: the highest published rates are for coding, STEM and professional expertise, and platforms such as Mercor and Alignerr recruit specialists. Skills that models cannot yet judge stay valuable longer.

Source: DataAnnotation: FAQs; TechCrunch: Mercor quintuples valuation to $10B with $350M Series C (Oct 27, 2025); Alignerr homepage (powered by Labelbox)

What are the long-term prospects for annotation work?

Expect a shift from volume to expertise. Generic tasks will pay less and become scarcer, while specialist evaluation in law, medicine, science, coding and languages continues to command higher rates. Use the work to build transferable evaluation and AI skills.

Source: DataAnnotation: FAQs

Does anyone else feel bad about participating in paid AI studies?

Yes, it is a common feeling. Some people resolve it by choosing projects they find meaningful, some by limiting how much they do, and some by leaving. You can decline projects that conflict with your values.

How do workers balance ethical concerns about AI with an immediate need for income?

Many accept the work while setting limits: declining projects that cross personal lines, favouring safety-focused work, and building other skills so the work stays a choice. Meeting basic needs is a legitimate priority.