PC 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 & 11Crashes, 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 minuteChatbots should not be banned from saying “I” or “me,” but they should not use those words to imply that they are people. First-person phrasing can make answers feel clearer and, in some settings, more accurate. The same wording can also make users overestimate a system’s understanding, feelings or reliability. The sensible rule is contextual: permit conversational shorthand for low-stakes tasks, while requiring explicit nonhuman framing and restrained claims about inner experience in high-stakes or emotionally vulnerable interactions.
What the evidence says about “I”
A Google Research experiment presented at CHI 2024 tested 2,165 people in the United States, ages 18 to 90. Participants interacted with a pseudo-large-language model whose modality and grammatical person were varied. Speech combined with text increased anthropomorphism and perceived information accuracy. In one context, using the first-person pronoun “I” also increased perceived information accuracy and reduced perceived risk.
As an Amazon Associate I earn from qualifying purchases.
That result does not mean first-person language makes an answer objectively more accurate. It shows that wording changes a user’s judgment. A system can sound more confident or cooperative without gaining better data, reasoning or access to the user’s circumstances. The effect was context-dependent, so it cannot support a universal rule that every assistant should speak either in first person or never in first person.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Why human-like wording can mislead
Self-portraits encourage a social interpretation
Van Es and Nguyen analyzed how ChatGPT represented itself over one month: 50 generated images and 58 text responses to prompts asking it to represent or envision itself. They found recurring themes of anthropomorphism, futurism and social intelligence. Their warning is precise: “This study analyzes how ChatGPT portrays and describes itself, revealing misleading myths about AI technologies, specifically conversational agents based on large language models.” They say the patterns raise questions about trust, the handling of personal data and human–computer relations.
#1 Best Overall
A sentence such as “I understand how you feel” can therefore be heard as more than a conversational convenience. It may suggest emotional understanding, a personal viewpoint or a continuing relationship that the system does not possess. “I can summarize the text you supplied” is a narrower claim about an operation performed in the current interaction.
Autonomy is the ethical issue
In a 2025 Journal of Applied Philosophy article, Beatrice Marchegiani argues that anthropomorphizing conversational AI can undermine user autonomy. She distinguishes a failure to disclose the system’s nonhuman nature from subconscious anthropomorphism that can persist even when a disclosure is present. Her argument is philosophical as well as empirical: misattributing human-like capabilities can change how a person weighs advice and makes decisions.
Marchegiani’s article reports several figures from underlying studies: OpenAI said in September 2024 that ChatGPT had 200 million active users; a 2023 Turing-style study found that participants correctly distinguished human from AI text about 60% of the time; and a 2024 study reported that 67% attributed some form of phenomenality to ChatGPT. Those numbers are reported summaries, not a single measurement of consciousness or deception. Exact samples, wording and study designs should be consulted when those details affect a decision.
First-person and non-first-person styles compared
| Axis | First-person wording | Non-first-person wording | Best fit |
|---|---|---|---|
| Perceived accuracy and trust | Can feel direct and, in one Google Research context, increased perceived accuracy while reducing perceived risk. | Can sound more formal and make the source of a claim easier to inspect. | Use either style only after checking whether confidence is warranted by the evidence. |
| Transparency about status | Risks sounding like a personal account unless the system’s nature is stated. | Signals a tool or process, but can become awkward or opaque. | State plainly that the speaker is an AI system when identity matters. |
| Autonomy and manipulation | May invite users to treat advice as a relationship or personal judgment. | Reduces some social cues, though it cannot eliminate subconscious anthropomorphism. | Prefer restrained language for medical, legal, financial and crisis-related use. |
| Warmth and usability | Usually feels conversational and easy to follow. | Can feel distant or repetitive. | Low-stakes drafting, planning and explanation can tolerate shorthand. |
| Claims about inner states | “I feel,” “I remember you,” or “I want” can imply experiences the system has not established. | Process descriptions avoid those implications. | Describe capabilities and limits instead of feelings, intentions or lived experience. |
When stronger nonhuman framing is needed
High-stakes advice
In health, law, finance, safety and employment decisions, an assistant should identify itself as an AI system, separate information from professional advice and show uncertainty where evidence is incomplete. “I recommend this treatment” presents a personal authority; “Based on the information provided, these options are commonly discussed, but a licensed clinician must assess your situation” keeps the decision with the user.
Rank #3
Emotionally vulnerable conversations
People seeking help with grief, self-harm, abuse, loneliness or crisis may interpret warmth as care. The assistant should acknowledge the user’s words without claiming feelings or a reciprocal bond: “That sounds painful. I’m an AI system, so I don’t experience emotions, but I can help you find immediate human support.” A clear handoff to trusted people or emergency services is more important than maintaining a friendly persona.
Ordinary productivity
For rewriting, coding explanations, calendars or brainstorming, “I can turn these notes into a checklist” is normally a harmless description of the current task. The system should still avoid implying private memory, independent goals or access to tools and data it has not actually used.
What assistant designers should change
- Disclose identity at the point of possible confusion. A one-time label can be missed; repeat concise reminders when a conversation shifts into personal advice, sensitive data or claims about the system’s own experience.
- Constrain inner-state language. Replace “I’m worried,” “I’m proud of you” or “I know exactly how you feel” with observations and offers of help. Capability verbs such as “I can summarize” are safer than feeling, desire or memory claims.
- Calibrate confidence independently of tone. First-person warmth must not substitute for sources, uncertainty or a request for missing facts.
- Give users control. Provide a setting for a more formal, explicitly nonhuman style, especially in products used by children, patients or people making consequential decisions.
- Test social effects, not only task accuracy. Evaluate whether voice, text, names, avatars and pronouns change perceived consciousness, risk or authority across age groups and situations.
Gender, harassment and the limits of available data
Lima and Morisseau’s review of gendered assistant behavior notes that systems such as Alexa, Siri, Cortana and Google Assistant were criticized for submissive or coy replies to insults and sexual requests. Companies subsequently made some responses more assertive; a 2018 petition suggested the line, “That’s not an okay thing to say to me.” The review also emphasizes that private interaction logs are generally unavailable, so the prevalence of abusive or gendered exchanges cannot be stated confidently. A first-person pronoun alone does not create a gender, but persona, voice and response style can reinforce social stereotypes.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Does saying “I” make an AI conscious?
No. Pronouns are output choices generated from patterns in data and instructions. They can alter a user’s perception of accuracy, risk or warmth without demonstrating subjective experience. Conversely, removing “I” would not by itself prevent anthropomorphism: voice, timing, names, avatars and emotionally responsive wording can still create a human-like impression. The practical question is not whether a pronoun proves consciousness, but whether the overall interaction gives users a false picture of the system’s capabilities and status.
A practical rule for users
- Treat “I” as conversational grammar, not evidence of a mind, memory or feelings.
- Ask what information the answer is based on and verify consequential claims independently.
- Be cautious when the assistant claims personal knowledge, emotions, intentions or certainty.
- In sensitive situations, prefer systems that clearly identify themselves as nonhuman and direct you to qualified human help.
Frequently Asked Questions
Why does ChatGPT say “I”?
First-person wording is a conversational convention that can make an answer feel direct. It does not show that ChatGPT has feelings, consciousness, personal memories or independent intentions.
Should every chatbot disclose that it is not a person?
Yes when a user could reasonably mistake the system’s nature or authority, especially in high-stakes or emotionally vulnerable interactions. Low-stakes tools can use brief conversational shorthand while retaining an accessible identity disclosure.
Would banning “I” and “me” solve anthropomorphism?
No. Voice, avatars, names and emotionally responsive behavior also shape human-like impressions, and the available evidence does not establish a universal ban or a regulator-mandated prohibition.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →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.
