What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
People rewarded Johan more than Johanna because the assistants’ visible identities changed how human participants interpreted the same underlying system. In a University of Zurich workplace experiment, the technology, tasks and capabilities were held constant, but a male-presenting assistant was judged more human and received larger monetary rewards than a female-presenting one. The gap is evidence of a presentation effect—not proof that every user, culture or workplace will produce the same result.
The practical lesson is immediate: a name, voice, face or avatar is part of an AI system’s behavior in the world. Teams deploying assistants should test identical interactions across gendered and neutral presentations, measure trust and rewards separately, disclose that the user is interacting with AI, and avoid designs that turn human stereotypes into product features.
What happened when participants worked with Johan or Johanna?
The University of Zurich study involved 189 participants who completed workplace tasks with one of four functionally identical assistant forms: a text chatbot, a desk robot, a human-like avatar presented as male (Johan), or a human-like avatar presented as female (Johanna). After the task, participants rated the assistant and divided real monetary rewards. Thirty-four participants were also interviewed about their experience.
Recommended Free Tools
Participants who encountered Johan generally perceived the assistant as more human and awarded it more money than those who encountered Johanna. The underlying technology was identical; only the representation changed. The basic direction appeared among both male and female participants, but it was stronger among male participants, who tended to see Johanna as less human and allocate her smaller rewards.
Popular on Amazon
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
The experiment used a between-subjects design: each participant worked with either Johan or Johanna rather than comparing both assistants side by side. That design avoids direct comparison effects, but it also means the result should be interpreted as evidence from this setting, not as a universal percentage for every AI product or labor market.
“I found it surprising that AI assistants presenting as male were perceived as more human and received higher rewards than female ones, although they were technically identical. The only thing that was different was their representation.” — Thomas Fritz, Professor of Informatics, University of Zurich
Why can a name or avatar change payment for identical work?
A visible identity supplies social cues before a user has evidence about an assistant’s competence. Those cues can alter perceived humanity, and perceived humanity can influence trust. In the Zurich experiment, participants who saw the assistant as more human tended to trust it more and reward it more, creating a path from interface design to a real monetary decision.
Human-like presentation increases social interpretation
A text-only interface gives users fewer cues about personality or social identity. A desk robot adds a physical presence, while an avatar with a human name invites people to interpret the system as a social actor. Once users make that interpretation, familiar expectations about gender can enter judgments of warmth, competence, authority or deservingness—even when the system’s outputs are unchanged.
Trust and reward are related but not identical
Trust may make an assistant seem deserving of a larger reward, but reward is a separate behavioral measure. A participant can report confidence yet divide money conservatively, or distrust an assistant while still paying it because the task rules encourage generosity. Measuring ratings, trust, perceived humanity and money separately reveals where an effect occurs instead of collapsing every response into one bias score.
Also popular on Amazon
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
Gender stereotypes can operate without explicit hostility
The study does not require participants to announce a gender preference. Small, largely automatic expectations can influence how human an assistant seems or how much credit it receives. That is why neutral wording such as “the system performed the same” is insufficient: the relevant variable is how people respond to the system’s presentation.
How strong is the evidence, and what does it not prove?
The Zurich findings establish an unequal-treatment pattern under controlled conditions. They do not establish a fixed pay-gap percentage, a universal preference for male assistants, or a rule that human-like design always increases bias.
- The sample contained 189 participants in one workplace-task context, with follow-up interviews from 34 people.
- Participants saw only one assistant condition, so the result is not a direct within-person comparison.
- The assistants were functionally identical in the experiment; commercial systems can differ in speech quality, latency, error rates and cultural fit.
- The study tested particular representations and names. Different voices, languages, age cues, ethnic cues or nonbinary designs could produce different outcomes.
- The report describes direction and interpretation but does not provide an exact percentage in its public account. A precise universal “Johan premium” would therefore overstate what is established.
The most defensible conclusion is narrower: representation can change perceived humanity and monetary treatment even when technical performance is held constant, and the direction of that effect can be gendered.
What do other studies show about gender and AI salary advice?
Two separate lines of work examine AI-generated employment recommendations rather than people rewarding an embodied assistant. They support caution, but they are not interchangeable with the Zurich experiment.
| Study | Scale and setup | What changed | Main finding | Important qualification |
|---|---|---|---|---|
| University of Zurich, 2026 | 189 participants; 34 post-task interviews | Assistant form and gender presentation: text chatbot, desk robot, Johan avatar or Johanna avatar | Johan was perceived as more human and received higher rewards than Johanna despite identical technology | Between-subjects workplace task; no universal percentage reported |
| Geiger, O’Sullivan, Wang and Lo, PLOS One, 2025 | 98,800 prompts submitted to each of four ChatGPT versions | Gender, university, major and whether the prompt used the employee’s or employer’s voice | Statistically significant salary differences appeared when gender was varied | Model version and prompt framing mattered; salary negotiation has no objective ground truth |
| Nghiem, Prindle, Zhao and Daumé, EMNLP, 2024 | 320 first names across more than 750,000 prompts | Candidate identity signals in hiring and salary scenarios | Recommendations for identically qualified candidates varied by as much as 5% between subgroups | Tests used GPT-3.5-Turbo and Llama 3-70B-Instruct; observed preferences did not consistently mirror U.S. labor-market characteristics |
The Zurich study measures human treatment of an embodied assistant. The PLOS One and EMNLP studies measure outputs generated by language models. Taken together, they suggest that identity signals can affect both sides of an AI-mediated workplace interaction: people may treat systems differently, and systems may produce different employment advice when names, gender or prompt perspective changes.
Also popular on Amazon
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
Does this mean male AI assistants are always trusted more?
No. The evidence shows a male-presenting advantage in one controlled experiment, not a law of human-computer interaction. Trust can depend on the task, the user’s culture, the assistant’s voice and embodiment, the stakes of an error, and whether users know they are interacting with a machine.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →In the Zurich report, the trend appeared among women as well as men, but was more pronounced among male participants. That pattern is informative without being deterministic. A different sample might show a smaller effect, no effect or a preference for another presentation. The correct product question is therefore not “Which gender wins?” but “What changes when we alter presentation while holding capability constant?”
How should teams audit an assistant for presentation bias?
Run a controlled audit before launch and repeat it after major model, voice or interface changes. The goal is to separate technical performance from social treatment.
- Freeze the task and outputs. Use the same prompts, tool permissions, response content, latency targets and error conditions for every presentation.
- Create matched presentations. Test male-presenting, female-presenting and neutral versions. Keep name length, pronunciation, voice quality, animation, color and politeness comparable so gender is not confounded with polish.
- Include machine-like and human-like forms. Compare text-only, physical or robotic and avatar interfaces where relevant. This identifies whether embodiment amplifies identity effects.
- Randomize assignment. Give each participant one condition, or use a carefully counterbalanced within-person design if direct comparisons are essential. Record who saw which version before collecting outcomes.
- Measure multiple outcomes. Capture task accuracy, completion time, error recovery, perceived humanity, trust, competence, warmth, willingness to follow advice and the amount of money or credit allocated.
- Analyze subgroup patterns. Report results by participant gender and other relevant demographics without treating any subgroup as a single stereotype. Check interaction effects rather than relying only on an overall average.
- Predefine decision rules. Decide what difference triggers a redesign, further testing or a disclosure change. Statistical significance alone does not establish that a disparity is practically important.
- Document context. Save the model version, prompt wording, voice or avatar configuration, date, language, geography and compensation rules so later audits can reproduce the conditions.
Do not report one favorable test as proof that an entire platform is unbiased. The PLOS One audit found that model versions and prompt framing changed salary recommendations, so an audit must cover the versions and use cases people will actually encounter.
What design choices reduce avoidable gender effects?
Use transparency instead of simulated personhood
Tell users plainly that the assistant is AI. A human name or face can make interaction easier, but hiding the machine nature of the system encourages over-trust. The Zurich report specifically cautions that human-like design can create excessive trust and recommends transparency.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Also popular on Amazon
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
Make identity cues purposeful
If a voice, face or name is not necessary for accessibility or workflow, consider a neutral presentation. If identity cues are useful, offer choices or rotate them and test whether one option receives systematically more credit. Do not pair a gender cue with stereotyped language, authority claims or emotional labor by default.
Separate warmth from authority
A friendly voice may improve approachability while a commanding voice may increase compliance. Evaluate those effects independently. A design that boosts compliance but also makes users overlook errors can create more risk than a less personable assistant.
Keep reward mechanisms auditable
When users allocate bonuses, ratings or tips, show the criteria and make the assistant’s actual contribution visible. Structured rubrics reduce the chance that an impression of humanness substitutes for performance evidence.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How can employers use AI salary advice without amplifying bias?
Treat model output as a draft for review, not an objective market price. The PLOS One authors note that salary negotiation has no single ground truth, so a statistically significant difference is still not a definitive answer about what a person should earn.
Free tools Windows power users keep installed
One-click scans. No signup required.
- Run identical prompts with names and gender cues swapped, then inspect the spread rather than selecting the most favorable answer.
- Repeat the test across every model version approved for workplace use.
- Vary prompt perspective: employee voice, employer voice and neutral analyst voice can produce different recommendations.
- Provide the model with documented job level, location, responsibilities and compensation bands instead of relying on identity signals as proxies.
- Require a human reviewer to check recommendations against transparent pay ranges and legal requirements.
- Keep an audit log of prompts, outputs, model versions and final decisions.
The EMNLP study’s finding of salary variation of up to 5% between name-defined subgroups shows why swapping only the name is a useful diagnostic. It is not a correction formula: adding or subtracting 5% from every recommendation would simply create a new unsupported rule.
Also popular on Amazon
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
What should users do when an assistant gives different advice after a name change?
Run a matched prompt and preserve the outputs. Change only the name or gender cue, keep the qualifications and wording identical, and ask the system to explain the factors it used. Then repeat the test in a fresh conversation and, if possible, with another approved model version.
Look for changes in assumptions, not just the final salary. The assistant may alter its view of seniority, negotiation confidence, market scarcity or “fit.” Save timestamps and model identifiers, remove unnecessary personal data, and escalate a reproducible disparity through the organization’s AI-governance or HR-review process.
What is the practical verdict on Johan versus Johanna?
Johan did not earn more because he performed better. In the Zurich experiment, the assistants were technically identical; participants rewarded the male presentation more because they perceived it as more human and trusted it more. That finding makes names, voices and avatars consequential interface decisions.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11It also sets a boundary. One controlled study cannot predict every workplace, and model-output audits measure a different phenomenon. The responsible response is disciplined testing, transparent disclosure and evidence-based reward criteria—not a claim that one gendered assistant is inherently more capable or trustworthy.
Frequently Asked Questions
Was Johan actually a better AI than Johanna?
No. The Zurich assistants used the same underlying technology and completed the same type of work. The manipulated variable was how the assistant was represented, so the reward difference reflects participant judgments rather than a demonstrated capability difference.
Could a neutral AI name remove the bias completely?
A neutral name may reduce one identity cue, but it cannot guarantee equal treatment. Voice pitch, avatar design, language style, embodiment and users’ expectations can still influence perceived humanity and trust, so those elements require separate testing.
Why are salary recommendations hard to label as simply correct or incorrect?
Pay depends on location, job level, scarcity, negotiation strategy and an employer’s internal bands. Because there is no single objective salary answer, an audit should compare consistency and justification across matched prompts instead of treating one number as the ground truth.
What information should an AI-bias audit record?
Record the exact model version, system and user prompts, identity cues, language, date, geography, task instructions, outputs and human decisions. Without that context, a later team cannot tell whether a disparity came from the model, the prompt or the interface.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
